From pmui at groundworkopensource.com Mon Feb 4 18:03:50 2008 From: pmui at groundworkopensource.com (Peter Mui) Date: Mon, 4 Feb 2008 18:03:50 -0800 Subject: BayLISA Monitoring SIG: Weds, Feb 13 2008, 7PM References: <200711150754.lAF7sc08014636@firewall.harker.com> Message-ID: <5898EE9B-4DBB-4E02-9CB7-E713BD08B43A@groundworkopensource.com> (Hi: You're invited to the BayLISA Monitoring SIG, Weds, Feb 13 2008, 7PM. See the meeting announcement pasted below: feel free to post it and/or forward it along to anyone else who might be interested. Many thanks, and hope to see you there! -Peter) ================================================= Monitoring SIG XIV: Ganglia Monitoring and Integration Thomas Stocking will present an overview of Ganglia and showcase the Ganglia Integration Module that adds innovative thresholding functions to the Ganglia monitoring system. Then Peter Loh will show some enhancements he's developed to extend ganglia with alarms and filtered console views. This SIG will be a kick-off to the Ganglia "Project in Residence" event Feb 28-29: more details at the meeting. What: BayLISA Monitoring SIG XIV: Ganglia Monitoring and Integration Who: Anyone interested in IT monitoring issues and tools (newbies particularly welcome!) When: Wednesday, Feb 13 2008, 7PM Where: GroundWork Open Source, 139 Townsend St., San Francisco How: 139 Townsend St. is very near AT&T Ballpark. It is one and a half blocks from the CalTrain Depot. Take the MUNI N, T or J trolley to 2nd and King (ballpark stop) or take the 30 or 45 bus (among others) crosstown. Free evening street parking can probably be found, and there are several fee-based parking garages around in case of parking difficulty. Cost: Free!! Presidents' Day pizza, carbonated and non-carbonated liquid refreshments, and healthy (and questionably healthy) snacks provided by GroundWork. We'll open up the doors at 6:30 or so and start the formal part of the meeting promptly at 7PM. RSVP (not necessary, but helpful): Peter Mui, pmui at groundworkopensource.com , 415-992-4573, www.groundworkopensource.com Can't attend in person? We'll try to broadcast live via WebEx: contact Peter if you want more info on this. ================================================= -------------- next part -------------- An HTML attachment was scrubbed... URL: From holland at guidancetech.com Tue Feb 5 19:46:38 2008 From: holland at guidancetech.com (Rich Holland) Date: Tue, 5 Feb 2008 22:46:38 -0500 Subject: domain transfer problem Message-ID: <0a2e01c86872$e2305d40$a69117c0$@com> A few years ago I registered my domain name with sharednic.com, which was cheap at the time ($15/year). Now I'd like to transfer it to GoDaddy, but they tell me I need an "authorization code" from my current registrar. I've been trying for more than 2 weeks to reach someone at sharednic.com to get this information, and they don't respond to email or phone calls. I've emailed every address I can find for them (web pages, whois database, emails they've sent me when I originally registered, etc) and called their contact number in the UK, which is answered by an answering machine. GoDady says they can't process the transfer without the code, even though I've used my control panel on sharednic.com to "unlock" my domain; I thought that would be all I need to do.. It seems like my current registrar is just trying to ignore me to force me to renew through them, but I don't want to do business with a company that has such lousy customer service. Anyone have any advice how to make this work and get to someone with (1) a US contact number that is (2) answered by a human being? Thanks! Rich Holland Principal Consultant Guidance Technologies, Inc. Cell: 913-645-1950 Fax: 913-273-0675 -------------- next part -------------- An HTML attachment was scrubbed... URL: From rick at linuxmafia.com Tue Feb 5 21:02:10 2008 From: rick at linuxmafia.com (Rick Moen) Date: Tue, 5 Feb 2008 21:02:10 -0800 Subject: domain transfer problem In-Reply-To: <0a2e01c86872$e2305d40$a69117c0$@com> References: <0a2e01c86872$e2305d40$a69117c0$@com> Message-ID: <20080206050209.GJ23798@linuxmafia.com> Quoting Rich Holland (holland at guidancetech.com): > A few years ago I registered my domain name with sharednic.com, which was > cheap at the time ($15/year). Now I'd like to transfer it to GoDaddy, but > they tell me I need an "authorization code" from my current registrar. Authorization Codes are also called Extensible Provisioning Protocol (EPP) Codes, and are required for some TLDs but not others. Here's some material about them: http://en.wikipedia.org/wiki/EPP_code http://www.wdbc.com/domain/transfer-authcode.cfm http://customersupport.networksolutions.com/article.php?id=161 Since sharednic.com is apparently an Enom reseller, maybe you need to escalate any lack of responsiveness from sharednic.com to them. From afife at untangle.com Tue Feb 5 22:45:48 2008 From: afife at untangle.com (Andrew Fife) Date: Tue, 5 Feb 2008 22:45:48 -0800 (PST) Subject: Bruce Perens @ BALUG (Feb 19th) Message-ID: <002601c8688b$f757d310$4301a8c0@Untangle.local> Howdy Folks: Bruce Perens will be giving a talk titled "Innovation Goes Public" at the Bay Area Linux Users Group (BALUG) on February 19th. (See the abstract below) If you'd like to come, please RSVP: RSVP at balug.org If you haven't been to BALUG in a while, this a great opportunity to check out what we're up to... and who knows you may just wind up eating dinner with Bruce Perens at your table. Upcoming 2008 speakers include: Feb 19th - Bruce Perens March 25th (New Date) - Mark Shuttleworth (Ubuntu/Canonical) April 15th - Eric Allman (Sendmail) May 20th - Jeremy Allison (Samba) June 17th - Andrew Morton (Linux kernel) July 15th - Mike Linksvayer (Creative Commons) So why not signup for BALUG's extremely low volume announce list: http://lists.balug.org/listinfo.cgi/balug-announce-balug.org Meeting Details... 6:30pm February 19th, 2008 Four Seas Restaurant 731 Grant Ave. San Francisco, CA 94108 PARKING: http://www.portsmouthsquaregarage.com/ Cost: The meetings are always free, but dinner is $13 ABSTRACT: "Innovation Goes Public" Presented by Bruce Perens, co-founder of the Open Source initiative insoftware. Open Source provides much of the software infrastructure for many of the world's largest companies and organizations: Merrill Lynch, Google, Pixar, Amazon, the City of New York, and probably you - although you might not know it. Innovative products like Linux, Firefox, and Apache are the market-leaders in their sectors, but there are tens of thousands of Open Source programs, used for just about everything. But the economics of Open Source are non-intuitive: how can you make money by giving software away? Why did IBM de-emphasize AIX, after spending Billions, in favor of Linux, the product of a loose collaboration of programmers that it can never control? How can the world's greatest city trust Open Source to help manage its jails? Bruce Perens will show how Open Source is often the most effective strategy for creating and utilizing new innovation. He will explain the economics of Open Source and how it works for profit-generating companies. His talk will be clear to beginners yet informative even for Open Source pros. About BALUG: BALUG is lively gathering of Linux users & free software enthusiasts that combines great food, community & intimate access to featured speakers. We meet in the bar of the Four Seas Restaurant from 6:30pm. At 7pm, we share a family-style Chinese dinner, which is followed by our guest speaker. BALUG Mailing list Policy: BALUG promises not to abuse other LUGs mailing lists. Our current policy is to make one monthly announcement on other Bay Area LUGs mailing lists. If you feel this is not appropriate for a particular list, please tell us which list and what you feel would be a more appropriate policy for that list. Please send feedback to balug-contact at balug.org. ---------------------------------------- Andrew Fife Untangle - Open Source Security Gateway download.untangle.com 650.425.3327 (O) 415.806.6028 (C) afife at untangle.com From lgj at usenix.org Thu Feb 7 10:36:49 2008 From: lgj at usenix.org (Lionel Garth Jones) Date: Thu, 07 Feb 2008 10:36:49 -0800 Subject: EVT '08 Call for Papers Message-ID: <47AB4FC1.6010403@usenix.org> --------------------------------------- 2008 USENIX/ACCURATE Electronic Voting Technology Workshop (EVT '08) July 28-29, 2008 San Jose, CA, USA Sponsored by USENIX: The Advanced Computing Systems Association, and ACCURATE: A Center for Correct, Usable, Reliable, Auditable, and Transparent Elections Refereed paper submissions due: March 28, 2008, 11:59 p.m. PDT Panel proposals due: May 2, 2008 http://www.usenix.org/evt08/cfpa ----------------------------------------- The Call for Papers for the 2008 USENIX/ACCURATE Electronic Voting Technology Workshop is now available. EVT '08 seeks to bring together researchers from a variety of disciplines, ranging from computer science and human-computer interaction experts through political scientists, legal experts, election administrators, and voting equipment vendors. EVT seeks to publish original research on important problems in all aspects of electronic voting. We welcome papers on voting topics including but not limited to: - Voter registration and pre-voting - Vote collection - Vote tabulation - Post-election auditing - Design, implementation, and evaluation of new voting technologies and protocols - Scientific evaluations of existing voting technologies - System testing methodologies - Deployment and lifecycle issues - Threat mitigation - Usability - Accessibility - Legal issues, including ADA, HAVA, intellectual property, and nondisclosure agreements on voting system evaluations - Issues with and evolution of voting technology standards The submission deadline for refereed papers is 11:59 p.m. PDT on Friday, March 28, 2008. Paper submission guidelines can be found at http://www.usenix.org/evt08/cfpa In addition to paper presentations, the workshop may include panel discussions with substantial time devoted to questions and answers. Proposals for panels should include a brief abstract, a list of possible panelists and their affiliations, and an indication of which of those panelists have confirmed participation. Panel proposals are due May 2, 2008. Please submit proposals to evt08panels at usenix.org. EVT '08 will be a two-day event, Monday, July 28, and Tuesday, July 29, 2008, co-located with the 17th USENIX Security Symposium in San Jose, California. We look forward to your submissions. David Dill, Stanford University Tadayoshi Kohno, University of Washington EVT '08 Program Chairs --------------------------------------- 2008 USENIX/ACCURATE Electronic Voting Technology Workshop (EVT '08) July 28-29, 2008 San Jose, CA, USA Sponsored by USENIX: The Advanced Computing Systems Association, and ACCURATE: A Center for Correct, Usable, Reliable, Auditable, and Transparent Elections Refereed paper submissions due: March 28, 2008, 11:59 p.m. PDT Panel proposals due: May 2, 2008 http://www.usenix.org/evt08/cfpa ----------------------------------------- From lgj at usenix.org Tue Feb 12 14:47:12 2008 From: lgj at usenix.org (Lionel Garth Jones) Date: Tue, 12 Feb 2008 14:47:12 -0800 Subject: USENIX LISA '08 Call For Papers Message-ID: <47B221F0.7090002@usenix.org> -------------------------------------------------------------------- Call for Papers LISA '08: 22nd Large Installation System Administration Conference November 9-14, 2008, San Diego CA, USA http://www.usenix.org/lisa08/cfpa Extended abstract and paper submissions due: May 8, 2008 Sponsored by USENIX and SAGE -------------------------------------------------------------------- Dear Colleague, The LISA '08 organizers invite you to contribute proposals for refereed papers, invited talks, and workshops, plus any ideas you have for Guru Is In sessions, the Hit the Ground Running track, Work-in-Progress Reports, posters, and training sessions. The theme for LISA '08 is "Real World System Administration." The Call for Participation with submission guidelines and sample topics can be found on the USENIX Web site at http://www.usenix.org/lisa08/cfpa Since 1987, the annual LISA conference has become the premier meeting place for professional system and network administrators. System administrators of all ranks, from novice to veteran, and of all specialties meet to exchange ideas, sharpen skills, learn new techniques, debate current issues, and mingle with colleagues and friends. The conference's diverse group of participants is matched by an equally broad spectrum of activities: * A training program for both beginners and experienced attendees covers many administrative topics, ranging from basic administrative procedures to using cutting-edge technologies. * Refereed papers present the latest developments and ideas related to system and network administration. * Workshops, invited talks, and panels discuss important and timely topics in depth and typically include lively and/or controversial debates and audience interaction. * Work-in-Progress Reports (WiPs) and poster sessions provide brief looks ahead to next year's innovations. GET INVOLVED! * Submit a draft paper or extended abstract proposal for a refereed paper. * Propose a tutorial topic. * Suggest an invited talk or panel discussion. * Share your experience by leading a Guru Is In session. * Create and lead a workshop. * Propose a short Hit the Ground Running presentation. * Present a Work-in-Progress Report (WiP) or submit a poster. * Organize a Birds-of-a-Feather (BoF) session. * Email an idea to the program chair: lisa08ideas at usenix.org We look forward to hearing from you! On behalf of the LISA '08 Organizers, Mario Obejas, Raytheon LISA '08 Program Chair lisa08chair at usenix.org ------------------------------------------------------------ IMPORTANT DATES Extended abstract and paper submissions due: May 8, 2008 Invited talk and workshop proposals due: May 20, 2008 Guru Is In & Hit the Ground Running proposals due: May 31, 2008 Notification to authors: Mid-June 2008 Poster proposals due, first round: July 16, 2008 Notification to poster presenters, first round: July 23, 2008 Final papers due: August 20, 2008 Poster proposals due, second round: Oct. 22, 2008 Notification to poster presenters, second round: Oct. 29, 2008 Submission guidelines and more information can be found at http://www.usenix.org/lisa08/cfpa ------------------------------------------------------------- From extasia at extasia.org Tue Feb 12 16:10:29 2008 From: extasia at extasia.org (David Alban) Date: Tue, 12 Feb 2008 16:10:29 -0800 Subject: wtf: hostid gives '0' as a hostid Message-ID: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> saw something today i've never seen before. on a linux box, /usr/bin/hostid printed '0' to stdout (plus a newline, of course). uh... isn't it supposed to be an eight digit hex number? i copied hostid from a box on which it was working correctly to /tmp on the "bad" box. when i ran /tmp/hostid it sill gave a hostid of 0. each of these boxes is *supposed* to be a clone at the hardware level. our environment depends on /usr/bin/hostid on each host working correctly. a cursory web search hasn't been enlightening. has anyone run into this before? if so, how did you fix it? -- Live in a world of your own, but always welcome visitors. From rick at linuxmafia.com Tue Feb 12 20:34:41 2008 From: rick at linuxmafia.com (Rick Moen) Date: Tue, 12 Feb 2008 20:34:41 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> Message-ID: <20080213043440.GM4958@linuxmafia.com> Quoting David Alban (extasia at extasia.org): > saw something today i've never seen before. on a linux box, > /usr/bin/hostid printed '0' to stdout (plus a newline, of course). > uh... isn't it supposed to be an eight digit hex number? I believe that (on Linux) it's a hex rendition of the first ethernet port's IP, so I'm guessing your Linux host has that set to zeroes. Note correspondence between my hostid value and the eth1 IP address (eth0 being disabled): $ /sbin/ifconfig eth1 Link encap:Ethernet HWaddr 00:D0:B7:93:31:0E inet addr:198.144.195.186 Bcast:198.144.195.191 Mask:255.255.255.248 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:3620757 errors:0 dropped:0 overruns:0 frame:0 TX packets:3959775 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:494644356 (471.7 MiB) TX bytes:2192414672 (2.0 GiB) Interrupt:11 Base address:0x1080 Memory:fa202000-fa202038 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:149902 errors:0 dropped:0 overruns:0 frame:0 TX packets:149902 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:81642189 (77.8 MiB) TX bytes:81642189 (77.8 MiB) $ hostid 90c6bac3 $ bc 1.06 Copyright 1991-1994, 1997, 1998, 2000 Free Software Foundation, Inc. This is free software with ABSOLUTELY NO WARRANTY. For details type `warranty'. obase=16 198 C6 144 90 195 C3 186 BA quit $ (Other *ixes use different algorithms.) -- Cheers, "Reality is not optional." Rick Moen -- Thomas Sowell rick at linuxmafia.com From cerise at armory.com Tue Feb 12 21:48:41 2008 From: cerise at armory.com (cerise at armory.com) Date: Tue, 12 Feb 2008 21:48:41 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> Message-ID: <20080213054841.GI25274@boogeyman> On Tue, Feb 12, 2008 at 04:10:29PM -0800, David Alban wrote: > saw something today i've never seen before. on a linux box, > /usr/bin/hostid printed '0' to stdout (plus a newline, of course). > uh... isn't it supposed to be an eight digit hex number? > > i copied hostid from a box on which it was working correctly to /tmp > on the "bad" box. when i ran /tmp/hostid it sill gave a hostid of 0. > each of these boxes is *supposed* to be a clone at the hardware level. > > our environment depends on /usr/bin/hostid on each host working > correctly. a cursory web search hasn't been enlightening. has anyone > run into this before? if so, how did you fix it? > > -- > Live in a world of your own, but always welcome visitors. hostid works by using the gethostid() POSIX call. In glibc, it's implemented as trying to read /etc/hostid. If that fails, then it gets the ip address, shifts it left 16 bits and ors that against the ip address shifted right 16 bits. The lower 32 bits is what hostid prints. hostid might be 0 because of your /etc/hostid. It might be 0 because of the IP address or the contents of /etc/hostid. (this is from coreutils 6.9 and glibc-2.7. YMMV.) -Phil/CERisE From extasia at extasia.org Wed Feb 13 13:23:18 2008 From: extasia at extasia.org (David Alban) Date: Wed, 13 Feb 2008 13:23:18 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080213043440.GM4958@linuxmafia.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> Message-ID: <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> On Feb 12, 2008 8:34 PM, Rick Moen wrote: > $ /sbin/ifconfig > eth1 Link encap:Ethernet HWaddr 00:D0:B7:93:31:0E > inet addr:198.144.195.186 Bcast:198.144.195.191 Mask:255.255.255.248 > UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 > RX packets:3620757 errors:0 dropped:0 overruns:0 frame:0 > TX packets:3959775 errors:0 dropped:0 overruns:0 carrier:0 > collisions:0 txqueuelen:1000 > RX bytes:494644356 (471.7 MiB) TX bytes:2192414672 (2.0 GiB) > Interrupt:11 Base address:0x1080 Memory:fa202000-fa202038 > $ hostid > 90c6bac3 > bc 1.06 > obase=16 > 198 > C6 > 144 > 90 > 195 > C3 > 186 > BA interesting. that means that a host whose primary ip addr is 10.0.0.5 may have the same hostid as lots of other hosts around the world. good thing we can't get to those other 10.0.0.5 machines... anyway, this is good. it means if i can get the first non-localhost ip address from ifconfig output, i can create my own hostid if the hostid command on the system is borken. thanks. -- Live in a world of your own, but always welcome visitors. From extasia at extasia.org Wed Feb 13 13:25:56 2008 From: extasia at extasia.org (David Alban) Date: Wed, 13 Feb 2008 13:25:56 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080213054841.GI25274@boogeyman> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213054841.GI25274@boogeyman> Message-ID: <4c714a9c0802131325p12cb4b4dl99c1fd55f95f35aa@mail.gmail.com> On Feb 12, 2008 9:48 PM, wrote: > hostid works by using the gethostid() POSIX call. In glibc, it's implemented > as trying to read /etc/hostid. If that fails, then it gets the ip address, > shifts it left 16 bits and ors that against the ip address shifted right 16 > bits. > > The lower 32 bits is what hostid prints. > > hostid might be 0 because of your /etc/hostid. It might be 0 because of the > IP address or the contents of /etc/hostid. it turns out we have some internal dns problems that may be affecting things on that one box. also, i think /etc/hostid is a bsd thing. i can't find it on any linux boxes. -- Live in a world of your own, but always welcome visitors. From cerise at armory.com Wed Feb 13 13:49:51 2008 From: cerise at armory.com (cerise at armory.com) Date: Wed, 13 Feb 2008 13:49:51 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <4c714a9c0802131325p12cb4b4dl99c1fd55f95f35aa@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213054841.GI25274@boogeyman> <4c714a9c0802131325p12cb4b4dl99c1fd55f95f35aa@mail.gmail.com> Message-ID: <20080213214951.GJ25274@boogeyman> On Wed, Feb 13, 2008 at 01:25:56PM -0800, David Alban wrote: > it turns out we have some internal dns problems that may be affecting > things on that one box. That would make sense. It tries to get the IP from the hostname for reasons I don't quite understand. That might be it. > also, i think /etc/hostid is a bsd thing. i can't find it on any linux boxes. In this case, the read of /etc/hostid is definitely there in libc. -Phil/CERisE From rick at linuxmafia.com Thu Feb 14 21:29:24 2008 From: rick at linuxmafia.com (Rick Moen) Date: Thu, 14 Feb 2008 21:29:24 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> Message-ID: <20080215052924.GS4958@linuxmafia.com> Quoting David Alban (extasia at extasia.org): > interesting. that means that a host whose primary ip addr is 10.0.0.5 > may have the same hostid as lots of other hosts around the world. > good thing we can't get to those other 10.0.0.5 machines... I know of no implications of the hostid value for networking. Maybe I'm wrong about this (you tell me), but I get the somewhat fuzzy impression that its main use is in licensing managers for proprietary *ix software. At least, that's what appears implied in many of the discussions I find. It seems most used on HP/UX and Solaris, and I see questions about its use and underlying algorithms on Linux only about twice a decade. (I frankly couldn't even remember how it worked, which is why I had to play around with ifconfig and bc.) From rayw at rayw.net Fri Feb 15 18:50:02 2008 From: rayw at rayw.net (Ray Wong) Date: Fri, 15 Feb 2008 18:50:02 -0800 Subject: So what's everyone using to push databases around these days? Message-ID: <20080216025002.GE27350@jmmredux.doomba.com> Hey, so with ancient reference ideas like DFS and google file system and multicast rsync, what's everyone using to copy decent sized directories of data around these days? I've got a site with a semi-typical data cycle (1 or more updates per day) filling a couple hundred gigs of space (MySQL in several database dirs). It goes to a few dozen distributed targets, let's constrain the parallelism requirement to between 5 and 50 hosts. What I'd really like is a multicast scp to just blast data to all targets (though some bwlimiting from ssh would be a nice option too), but something like mrsync is entirely too fragile for my liking, as the first target gets all data and yet is the one used for deciding what to send everywhere else when re-syncing (so it might be the only host with the files in place already, ensuring no one else gets the file). rdist over ssh is okayish, but the lack of a multicast option concerns me. I fully could handle doing a unicast recovery if a host missed out on a multicast copy, but without any multicast seems like I'd lose a lot of performance out of the data originator. And rsync, well, it's just painfully slow. :) So, what's anyone else doing? I haven't really felt like I was pushing any new ground since Postini or UltraDNS, I'll bet this is old hat to quite a few of you. Is there an obvious solution I'm missing? From rayw at rayw.net Fri Feb 15 19:42:58 2008 From: rayw at rayw.net (Ray Wong) Date: Fri, 15 Feb 2008 19:42:58 -0800 Subject: So what's everyone using to push databases around these days? In-Reply-To: <20080216025002.GE27350@jmmredux.doomba.com> References: <20080216025002.GE27350@jmmredux.doomba.com> Message-ID: <20080216034258.GF27350@jmmredux.doomba.com> Oh, the whole point of this was that it's just a big enough cycle that I can't rely on replication to be fast enough. :) On Fri, Feb 15, 2008 at 06:50:02PM -0800, Ray Wong wrote: > > > Hey, so with ancient reference ideas like DFS and google file system and > multicast rsync, what's everyone using to copy decent sized directories > of data around these days? > > I've got a site with a semi-typical data cycle (1 or more updates per > day) filling a couple hundred gigs of space (MySQL in several database > dirs). It goes to a few dozen distributed targets, let's constrain the > parallelism requirement to between 5 and 50 hosts. > > What I'd really like is a multicast scp to just blast data to > all targets (though some bwlimiting from ssh would be a nice option > too), but something like mrsync is entirely too fragile for my liking, > as the first target gets all data and yet is the one used for deciding > what to send everywhere else when re-syncing (so it might be the only > host with the files in place already, ensuring no one else gets the file). > > rdist over ssh is okayish, but the lack of a multicast option concerns > me. I fully could handle doing a unicast recovery if a host missed out > on a multicast copy, but without any multicast seems like I'd lose a lot > of performance out of the data originator. > > And rsync, well, it's just painfully slow. :) > > So, what's anyone else doing? I haven't really felt like I was pushing > any new ground since Postini or UltraDNS, I'll bet this is old hat to > quite a few of you. Is there an obvious solution I'm missing? From afife at untangle.com Sat Feb 16 09:56:34 2008 From: afife at untangle.com (Andrew Fife) Date: Sat, 16 Feb 2008 09:56:34 -0800 (PST) Subject: Massive Ubuntu Installfest for Schools (March 1st) Message-ID: <003101c870c5$5716a830$4301a8c0@Untangle.local> On Saturday March 1st, Untangle and the ACCRC are organizing a massive installfest for Bay Area schools. We are refurbishing hundreds of older/discarded computers with Ubuntu and donating them to Bay Area schools. We need your help from Linux users installing Ubuntu at the 4 locations, which are San Francisco, Berkeley, San Mateo & Marin County. Signup sheets for each location are here: http://wiki.untangle.com/index.php/Installfest If you can make the installfest, you can still help by driving participation by blogging about the event or voting for it on Digg or Slashdot. http://slashdot.org/bookmark.pl?url=http://www.untangle.com/installfest http://www.digg.com/submit?url=http://www.untangle.com/installfest&phase=2 More info on the installfest here: www.untangle.com/installfest Why the event is cool: Helps spread F/OSS (Ubuntu, Firefox, OpenOffice & more) Helps bridges the Digital Divide with underprivileged users Keeps toxic computer equipment out of landfills (Aprox 25,000 pounds) Is a cool community effort Thanks in advance for your help! -Andrew ---------------------------------------- Andrew Fife Untangle - Open Source Security Gateway download.untangle.com 650.425.3327 (O) 415.806.6028 (C) afife at untangle.com From extasia at extasia.org Sat Feb 16 10:00:02 2008 From: extasia at extasia.org (David Alban) Date: Sat, 16 Feb 2008 10:00:02 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080215052924.GS4958@linuxmafia.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> Message-ID: <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> On Feb 14, 2008 9:29 PM, Rick Moen wrote: > I know of no implications of the hostid value for networking. Maybe I'm > wrong about this (you tell me), but I get the somewhat fuzzy impression > that its main use is in licensing managers for proprietary *ix software. having a unique id for each host means that i can with confidence know that i've processed a particular host in a set of hosts. say i have host foo.bar.bat. say it has two ip addresses. say there are two additional cnames that point to it. say i have a config file for a program that allows a user to specify a hostname or ip address for a host. and that any host may appear multiple times in the config file. say it's very important that the host get processed only once, for some reason. if there are tasks in the config file for: foo foo.bar foo.bar.bat addl_cname_01.bar.bat addl_cname_02.bar.bat 10.1.2.3 10.2.3.4 which are all references to the host foo.bar.bat, then i the only way i know to collect all the tasks for foo.bar.bat is for the program to connect to each of these "host handles" and get the hostid. then it can create sets of tasks for each host based on hostid, and all of the tasks identified by the hostnames / ip addresses above will become a single set. i suppose i could use hostname instead of hostid, but now that i know on what hostid is based, it makes even more sense to use it (or something similar). it's possible someone could mess up hostnames and/or dns. but if a host can be reached via the network, it's hostid should be a unique identifier. so the value for me is not with regard to networking per se, but truly one of determining the uniqueness of a machine, despite the many "host handles" we can use to get to that machine. p.s. i'm starting to think that maybe a mac address on a host might be better yet than the output of hostid... -- Live in a world of your own, but always welcome visitors. From adrian.cockcroft at gmail.com Sat Feb 16 11:20:45 2008 From: adrian.cockcroft at gmail.com (adrian cockcroft) Date: Sat, 16 Feb 2008 11:20:45 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> Message-ID: <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> Uniqueness of a machine is a slippery concept. Its possible for a mac address to be assigned on a per-machine basis, and used for all NICs (older Sun machines definitely did this). Mac addresses can also be per-NIC, and multiple IP addresses can be on the same Nic. Cluster failover could also migrate the mac address between machines. Partitionable hardware such as Sun F15K class and IBM P-series are one box with a variable number of partitions, perhaps even sharing the same NIC. Then there is VMWare and its brethren so you have multiple virtual machines on the same hardware, possibly with the same Nic and hostid, but different OS instances. I don't think there is a 100% reliable answer unless you can constrain the deployed configurations to a known limited set of options. ...interesting times. Adrian On Feb 16, 2008 10:00 AM, David Alban wrote: > On Feb 14, 2008 9:29 PM, Rick Moen wrote: > > I know of no implications of the hostid value for networking. Maybe I'm > > wrong about this (you tell me), but I get the somewhat fuzzy impression > > that its main use is in licensing managers for proprietary *ix software. > > having a unique id for each host means that i can with confidence know > that i've processed a particular host in a set of hosts. say i have > host foo.bar.bat. say it has two ip addresses. say there are two > additional cnames that point to it. say i have a config file for a > program that allows a user to specify a hostname or ip address for a > host. and that any host may appear multiple times in the config file. > > say it's very important that the host get processed only once, for > some reason. if there are tasks in the config file for: > > foo > foo.bar > foo.bar.bat > addl_cname_01.bar.bat > addl_cname_02.bar.bat > 10.1.2.3 > 10.2.3.4 > > which are all references to the host foo.bar.bat, then i the only way > i know to collect all the tasks for foo.bar.bat is for the program to > connect to each of these "host handles" and get the hostid. then it > can create sets of tasks for each host based on hostid, and all of the > tasks identified by the hostnames / ip addresses above will become a > single set. > > i suppose i could use hostname instead of hostid, but now that i know > on what hostid is based, it makes even more sense to use it (or > something similar). it's possible someone could mess up hostnames > and/or dns. but if a host can be reached via the network, it's hostid > should be a unique identifier. > > so the value for me is not with regard to networking per se, but truly > one of determining the uniqueness of a machine, despite the many "host > handles" we can use to get to that machine. > > p.s. i'm starting to think that maybe a mac address on a host might > be better yet than the output of hostid... > > > -- > Live in a world of your own, but always welcome visitors. > From wingedpower at gmail.com Sat Feb 16 11:49:55 2008 From: wingedpower at gmail.com (Wing Wong) Date: Sat, 16 Feb 2008 11:49:55 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> Message-ID: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Regarding vmware mac addresses, those are typically generated and made to be as unique as possible. However, this can be overridden. But it is possible to ensure uniqueness of mac addresses for virtual machines using vmware vm(s). Ensuring uniqueness takes more than just finding something unique to the instance. This is where infrastructure external to the host comes into play to keep track of systems. This will probably require a means of generating a unique serial number for the host yourself and keeping track of it with informational tags/files on both the host itself as wells as in your management/inventory system. This way, if you lose track of a host, you can employ a "best guess" aproximation of finding your host again. Definitely an interesting problem... Wing On 2/16/08, adrian cockcroft wrote: > Uniqueness of a machine is a slippery concept. > > Its possible for a mac address to be assigned on a per-machine basis, > and used for all NICs (older Sun machines definitely did this). Mac > addresses can also be per-NIC, and multiple IP addresses can be on the > same Nic. Cluster failover could also migrate the mac address between > machines. > > Partitionable hardware such as Sun F15K class and IBM P-series are one > box with a variable number of partitions, perhaps even sharing the > same NIC. Then there is VMWare and its brethren so you have multiple > virtual machines on the same hardware, possibly with the same Nic and > hostid, but different OS instances. > > I don't think there is a 100% reliable answer unless you can constrain > the deployed configurations to a known limited set of options. > > ...interesting times. > Adrian > > On Feb 16, 2008 10:00 AM, David Alban wrote: > > On Feb 14, 2008 9:29 PM, Rick Moen wrote: > > > I know of no implications of the hostid value for networking. Maybe I'm > > > wrong about this (you tell me), but I get the somewhat fuzzy impression > > > that its main use is in licensing managers for proprietary *ix software. > > > > having a unique id for each host means that i can with confidence know > > that i've processed a particular host in a set of hosts. say i have > > host foo.bar.bat. say it has two ip addresses. say there are two > > additional cnames that point to it. say i have a config file for a > > program that allows a user to specify a hostname or ip address for a > > host. and that any host may appear multiple times in the config file. > > > > say it's very important that the host get processed only once, for > > some reason. if there are tasks in the config file for: > > > > foo > > foo.bar > > foo.bar.bat > > addl_cname_01.bar.bat > > addl_cname_02.bar.bat > > 10.1.2.3 > > 10.2.3.4 > > > > which are all references to the host foo.bar.bat, then i the only way > > i know to collect all the tasks for foo.bar.bat is for the program to > > connect to each of these "host handles" and get the hostid. then it > > can create sets of tasks for each host based on hostid, and all of the > > tasks identified by the hostnames / ip addresses above will become a > > single set. > > > > i suppose i could use hostname instead of hostid, but now that i know > > on what hostid is based, it makes even more sense to use it (or > > something similar). it's possible someone could mess up hostnames > > and/or dns. but if a host can be reached via the network, it's hostid > > should be a unique identifier. > > > > so the value for me is not with regard to networking per se, but truly > > one of determining the uniqueness of a machine, despite the many "host > > handles" we can use to get to that machine. > > > > p.s. i'm starting to think that maybe a mac address on a host might > > be better yet than the output of hostid... > > > > > > -- > > Live in a world of your own, but always welcome visitors. > > > -- Wing Wong wingedpower at gmail.com From rayw at rayw.net Sat Feb 16 19:11:02 2008 From: rayw at rayw.net (Ray Wong) Date: Sat, 16 Feb 2008 19:11:02 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> Message-ID: <20080217031102.GH27350@jmmredux.doomba.com> Indeed, fairly safe to say that any uniqueness is valid only insofar as the local site policy enforces it. I've used SSH host keys to check uniqueness at some sites, but I've also shared host keys amongst interchangable hosts to smooth failover scenarios (clearly not at the same site). I suppose since we're likely all perl hax0Rz it could make sense to make a hash of hash of hash index using several items and try to remember never to violate uniqueness on all of them at once. :) Ray On Sat, Feb 16, 2008 at 11:20:45AM -0800, adrian cockcroft wrote: > Uniqueness of a machine is a slippery concept. > > Its possible for a mac address to be assigned on a per-machine basis, > and used for all NICs (older Sun machines definitely did this). Mac > addresses can also be per-NIC, and multiple IP addresses can be on the > same Nic. Cluster failover could also migrate the mac address between > machines. > > Partitionable hardware such as Sun F15K class and IBM P-series are one > box with a variable number of partitions, perhaps even sharing the > same NIC. Then there is VMWare and its brethren so you have multiple > virtual machines on the same hardware, possibly with the same Nic and > hostid, but different OS instances. > > I don't think there is a 100% reliable answer unless you can constrain > the deployed configurations to a known limited set of options. > > ...interesting times. > Adrian > > On Feb 16, 2008 10:00 AM, David Alban wrote: > > On Feb 14, 2008 9:29 PM, Rick Moen wrote: > > > I know of no implications of the hostid value for networking. Maybe I'm > > > wrong about this (you tell me), but I get the somewhat fuzzy impression > > > that its main use is in licensing managers for proprietary *ix software. > > > > having a unique id for each host means that i can with confidence know > > that i've processed a particular host in a set of hosts. say i have > > host foo.bar.bat. say it has two ip addresses. say there are two > > additional cnames that point to it. say i have a config file for a > > program that allows a user to specify a hostname or ip address for a > > host. and that any host may appear multiple times in the config file. > > > > say it's very important that the host get processed only once, for > > some reason. if there are tasks in the config file for: > > > > foo > > foo.bar > > foo.bar.bat > > addl_cname_01.bar.bat > > addl_cname_02.bar.bat > > 10.1.2.3 > > 10.2.3.4 > > > > which are all references to the host foo.bar.bat, then i the only way > > i know to collect all the tasks for foo.bar.bat is for the program to > > connect to each of these "host handles" and get the hostid. then it > > can create sets of tasks for each host based on hostid, and all of the > > tasks identified by the hostnames / ip addresses above will become a > > single set. > > > > i suppose i could use hostname instead of hostid, but now that i know > > on what hostid is based, it makes even more sense to use it (or > > something similar). it's possible someone could mess up hostnames > > and/or dns. but if a host can be reached via the network, it's hostid > > should be a unique identifier. > > > > so the value for me is not with regard to networking per se, but truly > > one of determining the uniqueness of a machine, despite the many "host > > handles" we can use to get to that machine. > > > > p.s. i'm starting to think that maybe a mac address on a host might > > be better yet than the output of hostid... > > > > > > -- > > Live in a world of your own, but always welcome visitors. > > From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From rob.markovic at gmail.com Sat Feb 16 21:37:17 2008 From: rob.markovic at gmail.com (Robi) Date: Sat, 16 Feb 2008 21:37:17 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080217031102.GH27350@jmmredux.doomba.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <20080217031102.GH27350@jmmredux.doomba.com> Message-ID: <97a9d8c80802162137j67b3e732qc0ccafa2e72393dd@mail.gmail.com> It seems in Linux hostid is pretty useless, as you'll almost always get a value of 127.0.0.1 in network-endian hexadecimal. 007f 0100 f7 00 00 01 127 . 0 . 0 . 1 having VMs won't help much either, depending on the guest OS you're running, some will behave better than others based how they use their networking or pull data for hostid. pseudo-unique identifiers can be had from the SMBIOS/DMI data (UUID value). not sure which tools one would use to gather this though. -- Rob -------------- next part -------------- An HTML attachment was scrubbed... URL: From cerise at armory.com Sat Feb 16 22:12:01 2008 From: cerise at armory.com (cerise at armory.com) Date: Sat, 16 Feb 2008 22:12:01 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <97a9d8c80802162137j67b3e732qc0ccafa2e72393dd@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <20080217031102.GH27350@jmmredux.doomba.com> <97a9d8c80802162137j67b3e732qc0ccafa2e72393dd@mail.gmail.com> Message-ID: <20080217061201.GF13866@boogeyman> A couple of misconceptions from what I posted: 1) It does *not* go by MAC address. It goes by IP address. 2) It acquires the IP address by doing a DNS lookup of the hostname It gets the hostname from __gethostname(). It then does a DNS lookup of that hostname. It then takes the source address from the DNS lookup and mangles it by shifting left 16 bits and oring that with the source address shifted right 16 bits. In your case, Rob, your system believes it is something that resolves to 127.0.0.1. This might be because your hostname is set to localhost or it might be because your /etc/hosts specifies that 127.0.0.1 is the address of your hostname. -Phil/CERisE On Sat, Feb 16, 2008 at 09:37:17PM -0800, Robi wrote: > It seems in Linux hostid is pretty useless, as you'll almost always get a > value of > 127.0.0.1 in network-endian hexadecimal. > > 007f 0100 > > f7 00 00 01 > > 127 . 0 . 0 . 1 > > having VMs won't help much either, depending on the guest OS you're running, > some will behave better than others based how they use their networking or > pull data for hostid. > > pseudo-unique identifiers can be had from the SMBIOS/DMI data (UUID value). > not sure which tools one would use to gather this though. > > -- Rob From rob.markovic at gmail.com Sat Feb 16 23:27:36 2008 From: rob.markovic at gmail.com (Robi) Date: Sat, 16 Feb 2008 23:27:36 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080217061201.GF13866@boogeyman> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <20080217031102.GH27350@jmmredux.doomba.com> <97a9d8c80802162137j67b3e732qc0ccafa2e72393dd@mail.gmail.com> <20080217061201.GF13866@boogeyman> Message-ID: <97a9d8c80802162327n6ca9a9a7n624fd44e7db55af@mail.gmail.com> In Linux it seems to take the first network adapter, which is usually lo. But here's something a bit odd. I have a system with many IPs bound to lo:x where x is 0-100+. There's also eth0 with it's own aliasing, but in this case, hostid reports an IP that is bound to lo:4 for some reason. And to correct a typo, the f7 in the third line down should be 7f as from the 2nd line. Should read 7f 00 00 01 which then correctly makes 127.0.0.1. After glancing at the /etc/hosts file, the reason is the hostname is set to that IP. Or rather the IP is associated with that hostname. In any case, a poor "key" to base the security of some application on. Let's call it practical obscurity. Good work Phil, looks like that's what the hostid code does as you said. -- Rob > > 127.0.0.1 in network-endian hexadecimal. > > > > 007f 0100 > > > > f7 00 00 01 > > > > 127 . 0 . 0 . 1 > -------------- next part -------------- An HTML attachment was scrubbed... URL: From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From rick at linuxmafia.com Sun Feb 17 02:18:26 2008 From: rick at linuxmafia.com (Rick Moen) Date: Sun, 17 Feb 2008 02:18:26 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> Message-ID: <20080217101826.GG4958@linuxmafia.com> Quoting David Alban (extasia at extasia.org): > having a unique id for each host means that i can with confidence know > that i've processed a particular host in a set of hosts. say i have > host foo.bar.bat. Yes, the merit in that is obvious. However, based on the little I've seen, hostid doesn't appear to qualify as a guaranteed globally unique host ID -- definitely not on Linux, and I have my doubts about other *ixes, as well. You might be better off using, say, hashes of sshd host keys. (Some of your hosts might not run an sshd. You could fix that. Or not. ;-> ) -- Cheers, "Reality is not optional." Rick Moen -- Thomas Sowell rick at linuxmafia.com From rick at linuxmafia.com Sun Feb 17 02:21:34 2008 From: rick at linuxmafia.com (Rick Moen) Date: Sun, 17 Feb 2008 02:21:34 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <97a9d8c80802162137j67b3e732qc0ccafa2e72393dd@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <20080217031102.GH27350@jmmredux.doomba.com> <97a9d8c80802162137j67b3e732qc0ccafa2e72393dd@mail.gmail.com> Message-ID: <20080217102134.GH4958@linuxmafia.com> Quoting Robi (rob.markovic at gmail.com): > It seems in Linux hostid is pretty useless, as you'll almost always > get a value of 127.0.0.1 in network-endian hexadecimal. Well, for DHCP hosts and others that have the loopback line appearing first in /etc/hosts, anyway. Not for any other systems. From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From rayw at rayw.net Sun Feb 17 03:53:31 2008 From: rayw at rayw.net (Ray Wong) Date: Sun, 17 Feb 2008 03:53:31 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> Message-ID: <20080217115331.GI27350@jmmredux.doomba.com> So, getting back to your actual requirements, the point is that you want to be able to uniquely identify machines for an automated process. Would that be a fair statement? This implies you have some sort of automated management system intended. Something like CFengine handles all that, but supposing you want to have some sort of parallel system in place that can scale with a specific unique identifier. As discussed, hostid will always return a value, and it's one that can't by default be expected unique and that you wouldn't want to arbitrarily be changing all the time. Taking a look at Sam'l's idea of simply creating your own ID file (let's call it /etc/site/myID since we're about to change it from his idea), he readily acknowledge 2 obvious limits, that of being done manually, and concommitantly, lack of scalability. The obvious merging of these 2 is to create a tool that can automatically create /etc/site/myID, and a mechanism to notice when it's not there yet. Let's refer to the tool as /usr/local/sbin/createmyID. It's likely just going to be a shell script that creates some random string and checks some central registry (say, a mysql DB table with the ID as the primary key) for uniqueness there, then saves it to said registry and /etc/site/myID. (Note that of using a DB and primary key, success in doing an INSERT would combine the first 2 steps, leaving the write to the file dependent. Then all you need to do is ensure it gets called once for each host, and only once. In pseudo-cfengine terms: groups: have_unique_id = ( FileExists(/etc/site/myID)) copy: ${masterfiles}/inputs/usr/local/sbin/makemyID dest=/usr/local/sbin/makemyID server=${server} mode=755 type=checksum !have_unique_id:: create_id shellcommands: create_id:: "/usr/local/sbin/createmyID" Simple enough to manage it running once per host, and creates the ability for any subsequent task to simply use the grab /etc/site/myID and use that to decide if it's processed the task for that host or not. On Sat, Feb 16, 2008 at 10:00:02AM -0800, David Alban wrote: > On Feb 14, 2008 9:29 PM, Rick Moen wrote: > > I know of no implications of the hostid value for networking. Maybe I'm > > wrong about this (you tell me), but I get the somewhat fuzzy impression > > that its main use is in licensing managers for proprietary *ix software. > > having a unique id for each host means that i can with confidence know > that i've processed a particular host in a set of hosts. say i have > host foo.bar.bat. say it has two ip addresses. say there are two > additional cnames that point to it. say i have a config file for a > program that allows a user to specify a hostname or ip address for a > host. and that any host may appear multiple times in the config file. > > say it's very important that the host get processed only once, for > some reason. if there are tasks in the config file for: > > foo > foo.bar > foo.bar.bat > addl_cname_01.bar.bat > addl_cname_02.bar.bat > 10.1.2.3 > 10.2.3.4 > > which are all references to the host foo.bar.bat, then i the only way > i know to collect all the tasks for foo.bar.bat is for the program to > connect to each of these "host handles" and get the hostid. then it > can create sets of tasks for each host based on hostid, and all of the > tasks identified by the hostnames / ip addresses above will become a > single set. > > i suppose i could use hostname instead of hostid, but now that i know > on what hostid is based, it makes even more sense to use it (or > something similar). it's possible someone could mess up hostnames > and/or dns. but if a host can be reached via the network, it's hostid > should be a unique identifier. > > so the value for me is not with regard to networking per se, but truly > one of determining the uniqueness of a machine, despite the many "host > handles" we can use to get to that machine. > > p.s. i'm starting to think that maybe a mac address on a host might > be better yet than the output of hostid... > > -- > Live in a world of your own, but always welcome visitors. From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From jeff at drinktomi.com Sun Feb 17 15:10:03 2008 From: jeff at drinktomi.com (Jeff Younker) Date: Sun, 17 Feb 2008 15:10:03 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080217101826.GG4958@linuxmafia.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> Message-ID: On Feb 17, 2008, at 2:18 AM, Rick Moen wrote: > You might be better off using, say, hashes of sshd host keys. (Some > of > your hosts might not run an sshd. You could fix that. Or not. ;- > > ) Depending on your hardware vendor it may be possible to get a real unique ID. Dells have a software accessible serial number. The magic incantation is something like: dmidecode | grep Serial\ Number | head -n1. We used this at the last company I worked for. Other vendors may have similar systems. The deployment system we were building went from bare-metal to full deployment, so the system had to be able to identify a host before it could supply an appropriate OS. Things like network addresses and SSH keys were installed based on this determined identity. There are advantages to using an infrastructure specific identifier. The switch port that a machine lives on would be a good one. This way location determines function, and you can replace a system simply by unracking the old box, racking the new box, and turning it on. -jeff From samlb at am-cath.org Sat Feb 16 20:54:30 2008 From: samlb at am-cath.org (Sam'l B) Date: Sat, 16 Feb 2008 20:54:30 -0800 Subject: wtf: hostid gives '0' as a hostid In-Reply-To: <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <7097bd8c0802161149j5793fa6dv7f44a5bf1483db84@mail.gmail.com> Message-ID: <47B7BE06.8060606@am-cath.org> Really Dumb Solution: Create a text file, like "/etc/DumbID" on each of the machines. Put something unique in each text file. Look at "/etc/DumbID" when you log on to each machine and act accordingly. Why is it dumb? 1) Done by hand 2) Doesn't scale at all well Why is it maybe not-so-dumb? 1) You know exactly what to look for, and don't have to parse source files, or generate PERL hacks. 2) It can be self-documenting. Sam'l Bassett Wing Wong wrote: > Regarding vmware mac addresses, those are typically generated and made > to be as unique as possible. However, this can be overridden. But it > is possible to ensure uniqueness of mac addresses for virtual machines > using vmware vm(s). > > Ensuring uniqueness takes more than just finding something unique to > the instance. This is where infrastructure external to the host comes > into play to keep track of systems. This will probably require a means > of generating a unique serial number for the host yourself and keeping > track of it with informational tags/files on both the host itself as > wells as in your management/inventory system. > > This way, if you lose track of a host, you can employ a "best guess" > aproximation of finding your host again. > > Definitely an interesting problem... > > Wing > > On 2/16/08, adrian cockcroft wrote: >> Uniqueness of a machine is a slippery concept. >> >> Its possible for a mac address to be assigned on a per-machine basis, >> and used for all NICs (older Sun machines definitely did this). Mac >> addresses can also be per-NIC, and multiple IP addresses can be on the >> same Nic. Cluster failover could also migrate the mac address between >> machines. >> >> Partitionable hardware such as Sun F15K class and IBM P-series are one >> box with a variable number of partitions, perhaps even sharing the >> same NIC. Then there is VMWare and its brethren so you have multiple >> virtual machines on the same hardware, possibly with the same Nic and >> hostid, but different OS instances. >> >> I don't think there is a 100% reliable answer unless you can constrain >> the deployed configurations to a known limited set of options. >> >> ...interesting times. >> Adrian >> >> On Feb 16, 2008 10:00 AM, David Alban wrote: >>> On Feb 14, 2008 9:29 PM, Rick Moen wrote: >>>> I know of no implications of the hostid value for networking. Maybe I'm >>>> wrong about this (you tell me), but I get the somewhat fuzzy impression >>>> that its main use is in licensing managers for proprietary *ix software. >>> having a unique id for each host means that i can with confidence know >>> that i've processed a particular host in a set of hosts. say i have >>> host foo.bar.bat. say it has two ip addresses. say there are two >>> additional cnames that point to it. say i have a config file for a >>> program that allows a user to specify a hostname or ip address for a >>> host. and that any host may appear multiple times in the config file. >>> >>> say it's very important that the host get processed only once, for >>> some reason. if there are tasks in the config file for: >>> >>> foo >>> foo.bar >>> foo.bar.bat >>> addl_cname_01.bar.bat >>> addl_cname_02.bar.bat >>> 10.1.2.3 >>> 10.2.3.4 >>> >>> which are all references to the host foo.bar.bat, then i the only way >>> i know to collect all the tasks for foo.bar.bat is for the program to >>> connect to each of these "host handles" and get the hostid. then it >>> can create sets of tasks for each host based on hostid, and all of the >>> tasks identified by the hostnames / ip addresses above will become a >>> single set. >>> >>> i suppose i could use hostname instead of hostid, but now that i know >>> on what hostid is based, it makes even more sense to use it (or >>> something similar). it's possible someone could mess up hostnames >>> and/or dns. but if a host can be reached via the network, it's hostid >>> should be a unique identifier. >>> >>> so the value for me is not with regard to networking per se, but truly >>> one of determining the uniqueness of a machine, despite the many "host >>> handles" we can use to get to that machine. >>> >>> p.s. i'm starting to think that maybe a mac address on a host might >>> be better yet than the output of hostid... >>> >>> >>> -- >>> Live in a world of your own, but always welcome visitors. >>> > > From rayw at rayw.net Sun Feb 17 16:05:10 2008 From: rayw at rayw.net (Ray Wong) Date: Sun, 17 Feb 2008 16:05:10 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> Message-ID: <20080218000510.GM27350@jmmredux.doomba.com> I must say, this is the first time I can recall an applicable LISA topic discussion in quite a while. :) What's funny is how nothing's really changed, of course. Same sort of problem we were handling when we started being SAs. Different tools, same limiting issues. Anyway, back to the topic: On Sun, Feb 17, 2008 at 03:10:03PM -0800, Jeff Younker wrote: > Depending on your hardware vendor it may be possible to get a real > unique ID. Dells have a software accessible serial number. The > magic incantation is something like: > > dmidecode | grep Serial\ Number | head -n1. > > We used this at the last company I worked for. Other vendors may have > similar systems. Everyone does, the issue is that no one's consistent. If you're lucky enough to manage an environment with a sole exclusive h/w provider, you're in good shape. Otherwise, you've got different length IDs, and some vendors don't bother to set a serial, or set them all to the same (number of zeroes or whatever). Dell's great for that, other h/w suppliers that may already exist in many environments, not so much. Does, however, work great for new systems, and provide a useful reminder about planning ahead and establishing sane requirements. If you're fortunate enough to be arriving early enough in a company's life to choose the first vendor and establish those practices. In my experience, most SAs fall to pressure to get things done immediately and the trap of "we'll automate it later," "we'll get the specs and install system up once we ship this first batch of machines," etc. While I do like to try to get it "right the first time" when I set things up for new companies/new environments, my bread and butter has always been to come into companies that have realized they've built themselves into a corner of integration limitations and pick the pieces up. Most don't have budget to refresh all the heterogenous systems to bring to consistency, so an artificial consistency is more likely. I suspect most SAs see the same thing as there are more SAs than there are new companies, afaik. Or maybe there are a lot of new companies that keep hiring the same fools to mess things up to start ;) > There are advantages to using an infrastructure specific identifier. > The > switch port that a machine lives on would be a good one. This way > location determines function, and you can replace a system simply by > unracking the old box, racking the new box, and turning it on. easier said than done IME. Figuring out which port a machine is on relies on a lot of chained trusts between network gear, often of different vintages and different authentication options. Knowing what VLAN is a lot easier, and often works well enough, but again only with a planned environment (essentially, a dedicated VLAN per type of host/installation), though it does lend itself well to transition: as hosts are made compliant, bring them onto their VLAN. legacy installations get left in the swap VLAN, with a clear agreement in place that these machines are "best effort" support and will never be held to any SLA standards, and more importantly, will *always* be lower priority than the structured VLAN system, both new setup and maintenance. (yes, you still fix them when production goes wack. You just get to make a big deal out of how your team made up for the lack of support and negotiate some immediate form of reward for them, and announce how this shortsighted need has delayed ALL your teams deliverables by however long it took to fix the problem and follow-through) From rob.markovic at gmail.com Mon Feb 18 09:20:51 2008 From: rob.markovic at gmail.com (Robi) Date: Mon, 18 Feb 2008 09:20:51 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080218000510.GM27350@jmmredux.doomba.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> Message-ID: <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> Nice discussion. dmidecode | grep UUID should be a good provide a good unique id for similar hardware, and if it's very disimilar, then it's a good idea to hash it with something else to generate the final ID. Since there's a possibility of generating 2 of the same IDs if the UUID is somehow blank for 2 machines, adding a temporal factor to the hash will alleviate that (time of hash with ms). -- Rob -------------- next part -------------- An HTML attachment was scrubbed... URL: From asheesh at asheesh.org Mon Feb 18 12:22:26 2008 From: asheesh at asheesh.org (Asheesh Laroia) Date: Mon, 18 Feb 2008 12:22:26 -0800 (PST) Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> Message-ID: On Mon, 18 Feb 2008, Robi wrote: > Nice discussion. dmidecode | grep UUID should be a good provide a good > unique id for similar hardware, and if it's very disimilar, then it's a > good idea to hash it with something else to generate the final ID. Since > there's a possibility of generating 2 of the same IDs if the UUID is > somehow blank for 2 machines, adding a temporal factor to the hash will > alleviate that (time of hash with ms). My laptop says: UUID: Not Settable My suggestion would be, if you want to go totally nuts, dmidecode | md5sum. So long as the hardware doesn't change, that'd be a pretty good reference point. (Bonus points for grep -v'ing out the RAM or a few other volatile parts, I suppose.) If you dmidecode | hash_function > /etc/hostid once, when the machine is installed, that seems pretty good to me. -- Asheesh. -- Life is the childhood of our immortality. -- Goethe From cerise at armory.com Mon Feb 18 17:51:21 2008 From: cerise at armory.com (cerise at armory.com) Date: Mon, 18 Feb 2008 17:51:21 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> Message-ID: <20080219015121.GA28717@boogeyman> On Mon, Feb 18, 2008 at 12:22:26PM -0800, Asheesh Laroia wrote: > If you dmidecode | hash_function > /etc/hostid once, when the machine is > installed, that seems pretty good to me. > > -- Asheesh. Only one problem with that -- the result returned by hostid will be only 8 hex digits long. If you're going to go by that, then what's so wrong with using the IP address the way that hostid does? -Phil/CERisE From jeff at drinktomi.com Mon Feb 18 19:07:39 2008 From: jeff at drinktomi.com (Jeff Younker) Date: Mon, 18 Feb 2008 19:07:39 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080219015121.GA28717@boogeyman> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> <20080219015121.GA28717@boogeyman> Message-ID: <76E90D4C-79E7-48E7-9A64-1A242403868C@drinktomi.com> On Feb 18, 2008, at 5:51 PM, cerise at armory.com wrote: > Only one problem with that -- the result returned by hostid will be > only 8 > hex digits long. If you're going to go by that, then what's so > wrong with > using the IP address the way that hostid does? 1) IP addresses change. 2) when using dhcp all machines end up using 127.0.0.1 as the basis for their hostid. - Jeff Younker - jeff at drinktomi.com - From cerise at armory.com Mon Feb 18 19:29:51 2008 From: cerise at armory.com (cerise at armory.com) Date: Mon, 18 Feb 2008 19:29:51 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <76E90D4C-79E7-48E7-9A64-1A242403868C@drinktomi.com> References: <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> <20080219015121.GA28717@boogeyman> <76E90D4C-79E7-48E7-9A64-1A242403868C@drinktomi.com> Message-ID: <20080219032951.GB28717@boogeyman> On Mon, Feb 18, 2008 at 07:07:39PM -0800, Jeff Younker wrote: > 1) IP addresses change. Irrelevant, so long as /etc/hostid changes with it. It marks a unique host. > 2) when using dhcp all machines end up using 127.0.0.1 as the > basis for their hostid. That's not true. All machines that have their hostname resolve to 127.0.0.1 use 127.0.0.1 as the basis for their hostid. Using the -c switch with dhcpcd could solve this problem pretty easily. -Phil/CERisE From cerise at armory.com Mon Feb 18 19:38:35 2008 From: cerise at armory.com (cerise at armory.com) Date: Mon, 18 Feb 2008 19:38:35 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080219032951.GB28717@boogeyman> References: <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> <20080219015121.GA28717@boogeyman> <76E90D4C-79E7-48E7-9A64-1A242403868C@drinktomi.com> <20080219032951.GB28717@boogeyman> Message-ID: <20080219033835.GC28717@boogeyman> I should add too that I don't know of a DHCP server which will give a different address to a system renewing its lease, nor can I think of a particularly compelling reason for one to do so. -Phil/CERisE On Mon, Feb 18, 2008 at 07:29:51PM -0800, cerise at armory.com wrote: > On Mon, Feb 18, 2008 at 07:07:39PM -0800, Jeff Younker wrote: > > 1) IP addresses change. > > Irrelevant, so long as /etc/hostid changes with it. It marks a > unique host. > > > 2) when using dhcp all machines end up using 127.0.0.1 as the > > basis for their hostid. > > That's not true. All machines that have their hostname resolve > to 127.0.0.1 use 127.0.0.1 as the basis for their hostid. > > Using the -c switch with dhcpcd could solve this problem pretty > easily. > > -Phil/CERisE From jeff at drinktomi.com Mon Feb 18 20:21:02 2008 From: jeff at drinktomi.com (Jeff Younker) Date: Mon, 18 Feb 2008 20:21:02 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080219032951.GB28717@boogeyman> References: <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> <20080219015121.GA28717@boogeyman> <76E90D4C-79E7-48E7-9A64-1A242403868C@drinktomi.com> <20080219032951.GB28717@boogeyman> Message-ID: <08F49AFA-72A0-4DBA-A1A0-B4D53F6DC0A8@drinktomi.com> On Feb 18, 2008, at 7:29 PM, cerise at armory.com wrote: > On Mon, Feb 18, 2008 at 07:07:39PM -0800, Jeff Younker wrote: >> 1) IP addresses change. > > Irrelevant, so long as /etc/hostid changes with it. It marks a > unique host. That assumes that your deployment system works with hosts that have an OS installed. A deployment system that I helped to build had to identify systems before this was available. When unknown systems were placed onto the network they were given a minimal OS install by DHCP and placed into a 'limbo' network consisting of unallocated machines. This minimal OS install contacted a central server (identified by a bootp parameter) and fed over its identifying information. This included a serial number (dell hardware) and MAC address along with as much various data about its hardware configuration. The new information had two effects. Either it created a new row in a hosts database, or it updated an existing record with the mac address. If the information in the database was sufficient to assign the host to a cluster then a new DHCP configuration would be generated and DNS would be updated if needed. Finally the minimal OS would detect the change in assignment by querying the config server, and then reboot, this time coming up with a full OS install and its new identity. The upshot of this process is that it was never necessary to enter any networking information about a host. Networking ranges were assigned to clusters, and the install system determined the mac addresses on its own. The whole thing is possible because the system can locate a hardware identifying key. -jeff From jeff at drinktomi.com Mon Feb 18 20:32:28 2008 From: jeff at drinktomi.com (Jeff Younker) Date: Mon, 18 Feb 2008 20:32:28 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> Message-ID: On Feb 18, 2008, at 12:22 PM, Asheesh Laroia wrote: > My suggestion would be, if you want to go totally nuts, dmidecode | > md5sum. So long as the hardware doesn't change, that'd be a pretty > good reference point. (Bonus points for grep -v'ing out the RAM or > a few other volatile parts, I suppose.) That's actually a pretty cool idea, particularly with regard to pulling out the volatiles. > If you dmidecode | hash_function > /etc/hostid once, when the > machine is installed, that seems pretty good to me. Yeah, but if you've got the dmidecode function on the local system you don't really need /etc/hostid any more. -jeff From jeff at drinktomi.com Mon Feb 18 20:21:02 2008 From: jeff at drinktomi.com (Jeff Younker) Date: Mon, 18 Feb 2008 20:21:02 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080219032951.GB28717@boogeyman> References: <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> <97a9d8c80802180920i41cdf3a9rff78261486fb7c6d@mail.gmail.com> <20080219015121.GA28717@boogeyman> <76E90D4C-79E7-48E7-9A64-1A242403868C@drinktomi.com> <20080219032951.GB28717@boogeyman> Message-ID: <08F49AFA-72A0-4DBA-A1A0-B4D53F6DC0A8@drinktomi.com> On Feb 18, 2008, at 7:29 PM, cerise at armory.com wrote: > On Mon, Feb 18, 2008 at 07:07:39PM -0800, Jeff Younker wrote: >> 1) IP addresses change. > > Irrelevant, so long as /etc/hostid changes with it. It marks a > unique host. That assumes that your deployment system works with hosts that have an OS installed. A deployment system that I helped to build had to identify systems before this was available. When unknown systems were placed onto the network they were given a minimal OS install by DHCP and placed into a 'limbo' network consisting of unallocated machines. This minimal OS install contacted a central server (identified by a bootp parameter) and fed over its identifying information. This included a serial number (dell hardware) and MAC address along with as much various data about its hardware configuration. The new information had two effects. Either it created a new row in a hosts database, or it updated an existing record with the mac address. If the information in the database was sufficient to assign the host to a cluster then a new DHCP configuration would be generated and DNS would be updated if needed. Finally the minimal OS would detect the change in assignment by querying the config server, and then reboot, this time coming up with a full OS install and its new identity. The upshot of this process is that it was never necessary to enter any networking information about a host. Networking ranges were assigned to clusters, and the install system determined the mac addresses on its own. The whole thing is possible because the system can locate a hardware identifying key. -jeff From jeff at drinktomi.com Mon Feb 18 20:49:32 2008 From: jeff at drinktomi.com (Jeff Younker) Date: Mon, 18 Feb 2008 20:49:32 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080218000510.GM27350@jmmredux.doomba.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <20080217101826.GG4958@linuxmafia.com> <20080218000510.GM27350@jmmredux.doomba.com> Message-ID: <62B9475B-2B2D-4AE3-AFC6-2618AA55C265@drinktomi.com> > >> There are advantages to using an infrastructure specific identifier. >> The >> switch port that a machine lives on would be a good one. This way >> location determines function, and you can replace a system simply by >> unracking the old box, racking the new box, and turning it on. > > easier said than done IME. Figuring out which port a machine is on > relies > on a lot of chained trusts between network gear, often of different > vintages > and different authentication options. Acknowledged that it's not a simple problem, but it's not an unsolvable problem. It definitely requires cooperation from the networking folks too. > Knowing what VLAN is a lot easier, and often works well enough, but > again > only with a planned environment (essentially, a dedicated VLAN per > type of > host/installation), though it does lend itself well to transition: > as hosts > ... > deliverables by however long it took to fix the problem and follow- > through) That's a good approach, but it doesn't let you do things like distinguish boards within a blade server. In all of this you have to have a planned environment, and the advantages of maintaining order have to be such that people see the immediate benefit. (The path of most benefit must also be the path of least resistance.) -jeff From extasia at extasia.org Tue Feb 19 12:05:28 2008 From: extasia at extasia.org (David Alban) Date: Tue, 19 Feb 2008 12:05:28 -0800 Subject: [baylisa] Re: wtf: hostid gives '0' as a hostid In-Reply-To: <20080217031102.GH27350@jmmredux.doomba.com> References: <4c714a9c0802121610h5ef3e009nff4720d88fe6a26b@mail.gmail.com> <20080213043440.GM4958@linuxmafia.com> <4c714a9c0802131323y16e1078h2202e50d43ef22c6@mail.gmail.com> <20080215052924.GS4958@linuxmafia.com> <4c714a9c0802161000oa3eced4n676ec25739c6d8f5@mail.gmail.com> <3d9479a40802161120y161d6837ub6ca78c8ec999276@mail.gmail.com> <20080217031102.GH27350@jmmredux.doomba.com> Message-ID: <4c714a9c0802191205k149dfcb6y79300ba40fe30566@mail.gmail.com> On Feb 16, 2008 7:11 PM, Ray Wong wrote: > Indeed, fairly safe to say that any uniqueness is valid only insofar as the > local site policy enforces it. I've used SSH host keys to check uniqueness > at some sites, but I've also shared host keys amongst interchangable hosts > to smooth failover scenarios (clearly not at the same site). > > I suppose since we're likely all perl hax0Rz it could make sense to make > a hash of hash of hash index using several items and try to remember never > to violate uniqueness on all of them at once. :) i really liked the idea of using ssh host keys, but when i used ssh-keygen to look at the host key fingerprints of several of the candidate machines i found some of them sharing the same host key. so here's a working solution for me. it's not a universal solution. i have the advantage of having all linux machines in my problem space. it does an md5sum on the combined output of 'uname -a', the lines in ifconfig output that contain either an IP or mac address (i'm not bothering to filter out 127.0.0.1, although someone else could if they wanted to), and only the values in free output indicating total memory and swap on the machine. #/bin/bash export PATH=/sbin:/bin:/usr/sbin:/usr/bin hostname="$( hostname )" problems_encountered=0 uname_output="$( uname -a )" if [[ $? != 0 ]]; then echo 1>&2 "[$hostname] uname failed" problems_encountered=$(( $problems_encountered + 1 )) elif ! echo "$uname_output" | grep '^Linux' >/dev/null ; then echo 1>&2 "[$hostname] os type failed (not linux)" problems_encountered=$(( $problems_encountered + 1 )) fi ifconfig_output="$( ifconfig )" if [[ $? != 0 ]] ; then echo 1>&2 "[$hostname] ifconfig failed" problems_encountered=$(( $problems_encountered + 1 )) fi free_output="$( free )" if [[ $? != 0 ]] ; then echo 1>&2 "[$hostname] free failed" problems_encountered=$(( $problems_encountered + 1 )) fi { echo "$uname_output" echo "$ifconfig_output" | egrep '(inet|HWaddr) ' echo "$free_output" | awk '{ print $1,$2 }' | egrep '^(Mem|Swap):' } | md5sum | awk '{ print $1 }' exit $problems_encountered output is an md5sum. output of the echo block which is being summed would look something like: Linux some.host.name 2.6.9-55.0.2.ELsmp #1 SMP Tue Jun 12 17:59:08 EDT 2007 i686 i686 i386 GNU/Linux eth0 Link encap:Ethernet HWaddr 01:23:45:67:89:AB inet addr:10.11.12.13 Bcast:10.11.12.255 Mask:255.255.255.0 inet addr:127.0.0.1 Mask:255.0.0.0 Mem: 4152916 Swap: 2096472 for my solution, i only need the unique id to be good for about an hour. so if a host "has work done on it" and some of the attributes change, that's ok with me. i don't need data any older than an hour. -- Live in a world of your own, but always welcome visitors. From afife at untangle.com Wed Feb 27 15:11:50 2008 From: afife at untangle.com (Andrew Fife) Date: Wed, 27 Feb 2008 15:11:50 -0800 (PST) Subject: Donating 500 F/OSS Computers to Schools... SATURDAY!! (Pizza & Schwag from Mozilla!!) Message-ID: <02bc01c87996$396477c0$0200a8c0@Untangle.local> We're trying to donate 500 Ubuntu computers to Bay Area schools this SATURDAY... and we need your help! Can you lend a hand by volunteering to install Ubuntu? Signup to volunteer at one of the four locations (San Francisco, San Mateo, Berkeley or Novato) here: http://wiki.untangle.com/index.php/Installfest Untangle and the ACCRC refurbishing hundreds of older/discarded computers that the ACCRC has collected with Ubuntu and donating them to Northern California schools. We need support from the F/OSS community to help with the installs. We have automated as much of the install as possible so anyone can help regardless of their experience level. Of course we'll get error messages so we need gurus also! ...And the Mozilla Foundation is bringing pizza & schwag to each location! Thanks Mozilla! Installfest for Schools homepage: http://www.untangle.com/installfest Can't make the installfest but still want to help???? Help get the word out by blogging about it, Digg it or Slashdot it: http://slashdot.org/bookmark.pl?url=http://www.untangle.com/installfest http://www.digg.com/submit?url=http://www.untangle.com/installfest&phase=2 Why the event is cool: -Helps spread F/OSS (Ubuntu, Firefox, OpenOffice & more) -Helps bridges the Digital Divide with underprivileged users -Keeps toxic computer equipment out of landfills (Aprox 25,000 pounds) -Is a cool community effort Thanks in advance for your help! -Andrew ---------------------------------------- Andrew Fife Untangle - Open Source Security Gateway download.untangle.com 650.425.3327 (O) 415.806.6028 (C) afife at untangle.com From bill at wards.net Thu Feb 28 17:37:53 2008 From: bill at wards.net (bill at wards.net) Date: Thu, 28 Feb 2008 17:37:53 -0800 Subject: TONIGHT: PenLUG meeting 02/28/2008 Message-ID: +---------------------------------------------------------------+ |Date: |Thursday, February 28th, 2007 | |---------+-----------------------------------------------------| |Time: |meeting 7:00 - 9:00 PM, social/networking until 10 PM| |---------+-----------------------------------------------------| | |Bayshore Technology Park | |Location:|1300 Island Drive | | |Redwood City, CA 94065 | | |Suite 106 - Training Room | +---------------------------------------------------------------+ Thomas Belote, Wireless Mesh Networking As a grad student at SJSU, Thomas has worked on Wireless Mesh Networking and Mobile Ad Hoc Networking. The talk will compare solutions like OLSR and Microsoft's sorta open source Mesh Connectivity Layer (though it doesn't run on Linux). He will discuss why WDS is not sufficient and a mesh protocol is needed, and discuss the lack of openness thus far in 802.11s even though it is included in the OLPC, as well as security issues that I am still currently researching. RSVP Although it is not required, we like to have an idea of how many people to expect, so if possible please email rsvp at penlug.org if you are planning to attend. GETTING THERE For information on getting to the meeting, please see: http://maps.google.com/maps?q=1300+Island+Drive,+Redwood+City,+CA http://www.penlug.org/twiki/bin/view/Home/DrivingDirectionsQualys http://www.penlug.org/twiki/bin/view/Home/TransitDirectionsQualys CARPOOLING & TRANSIT Traffic on 101 can be pretty bad in the evening, so we encourage you to check traffic conditions before driving by dialing 5-1-1 on your phone or visiting www.511.org, and if possible to take public transit (best bet: bicycle via Caltrain) or carpool to this meeting. MORE INFORMATION See www.penlug.org for more information.