From alvin at Mail.Linux-Consulting.com Wed Jun 1 01:47:26 2005 From: alvin at Mail.Linux-Consulting.com (Alvin Oga) Date: Wed, 1 Jun 2005 01:47:26 -0700 (PDT) Subject: network thruput In-Reply-To: <1117597964.971.235367775@webmail.messagingengine.com> Message-ID: hi ya eric On Tue, 31 May 2005, Eric Wagar wrote: > > - current upper management premise ... > > - keep windoze and backup machines on the same lan > > because it doesn't clutter the wire > > even thou 80% of the packets is udp and netbios :-) duuhhh... > > Did you draw pretty pictures to dissuade them? That usually works for > me...the more colors the better. and a little script that anybody can run to show that transfering files on a network free of netbios is 5x faster but something musta been wrong with the test script..etc.etc.. it can't be .. etc..etc.. > So, all machines have a minimum of two NICs? some machines have 4 nics.. all going to the same nortel switch ( same 1u box ) and connected via is stackable cable in the rear some are channel bonded .. most are not > With your above idea, why not have the backup server on both LANs? backups of 500GB during the day kills the network .. which it does when it goes to the $150K 10-tape library thingie which in itself should be on a separate tape lan in my world, backups should be sorta invisible on the network, and see if the disk latency is tolerable for head seeks --- just thought it was fun ... "upper-level managers are always right" andits their $$$ and budget etc.. c ya alvin From michael at halligan.org Wed Jun 1 08:15:31 2005 From: michael at halligan.org (Michael T. Halligan) Date: Wed, 01 Jun 2005 08:15:31 -0700 Subject: network thruput In-Reply-To: References: Message-ID: <429DD113.6090109@halligan.org> >and a little script that anybody can run to show that transfering >files on a network free of netbios is 5x faster but something >musta been wrong with the test script..etc.etc.. it can't be .. etc..etc.. > > > >>So, all machines have a minimum of two NICs? >> >> > >some machines have 4 nics.. all going to the same nortel switch >( same 1u box ) and connected via is stackable cable in the rear > >some are channel bonded .. most are not > > One quick suggestion. quadruple your # of switches. I've built beowulf clusters where each server has 4-8 nics, each connected to a different switch, and then bonded, creating a 400 or 800mb (in your case 4gb or 8gb) network. If you have the spare hardware, it might be worth staging that over a weekend to see if there is an improvement. If there is, then that might not say that you were maxing out bandwidth, but it might say that you were maxing out the cpus on your switches. When you say they're getting a paltry 5-10MB/s at best, are you saying all of the servers are at the same time, or at any given time? Beyond that, the stacked switch setup could be bad if that means switch 10 has to traverse all of the other switches in order to get to switch 9. If that's the case, you might actually want to google for "hypertorroidal mesh" or "beowulf mesh" which will describe a couple of ways to build a network with multiple nics to lower the # of switching hops from any server to a percentage of any other server. Another thing I'd do is collect some good stats to show to the PHB's .. Setup NTOP for a week and show them that it's windows chatter eating up all the bandwidth. If they're manageable switches, setup cacti to graph them via snmp. Might also be worth digging in to see if you're having any type of arp or broadcast storms, perhaps a screwed up vlan. > > > >>With your above idea, why not have the backup server on both LANs? >> >> > >backups of 500GB during the day kills the network .. which it >does when it goes to the $150K 10-tape library thingie > which in itself should be on a separate tape lan > > > $150k? Ouch. For $20k nowadays you can get a 40 tape lto2 library that has 200GB (uncompressed) tapes.. Another $5-$7k you can go with the latest SDLT that's 300GB uncompressed (which I'm looking at purchasing this week, hence the numbers in my head) >in my world, backups should be sorta invisible on the network, >and see if the disk latency is tolerable for head seeks > > > I'm starting to give up on Tape to be honest. The value of tape and disk always goes back and fourth, but for a new infrastructure, where you're not in-bed with a backup software, I'm not sure if tape is always the right solution, excepts perhaps as an archive. I'm considering putting a small san in each of my datacenters, then coming up with an alternated backup scheme for all of our servers, and taking an archive of a snapshot every other day, maybe just once a week. Disk is just so fast, and getting so cheap.. It's like $10k for a 6 tb iscsi raid from apple, that's pretty appealing, especially with a good volume manager. >--- > >just thought it was fun ... "upper-level managers are always right" >andits their $$$ and budget etc.. > >c ya >alvin > > > -- ------------------- BitPusher, LLC http://www.bitpusher.com/ 1.888.9PUSHER (415) 724.7998 - Mobile From david.carmean at netapp.com Wed Jun 1 09:48:16 2005 From: david.carmean at netapp.com (David Carmean) Date: Wed, 1 Jun 2005 09:48:16 -0700 Subject: Recommendation for FreeBSD virtual private server vendor? Message-ID: <20050601164816.GO27610@netapp.com> Feeling guilty about the spam and diskspace load I'm putting on my friends' home/DSL server, I'm looking into affordable alternatives and Virtual Private Servers are looking interesting. I prefer FreeBSD over Linux. What I'm concerned about is how to get one set up with packet filters (pref. ip_filter) before the thing goes on-line. How do vendors handle that chicken-and-egg problem? Any recommendations for vendors or what to look for in services? RackSpace's DDS looks interesting. Is anybody reselling VMWare-based virtual machines? Thanks. -- David Carmean Network Appliance, Inc +1-408-822-6565 (ph) 495 E. Java Drive +1-408-822-4577 (fax) Sunnyvale, CA 94089 From eric at deadhookers.org Wed Jun 1 11:52:50 2005 From: eric at deadhookers.org (Eric Wagar) Date: Wed, 01 Jun 2005 11:52:50 -0700 Subject: network thruput In-Reply-To: References: Message-ID: <429E0402.5030308@deadhookers.org> >>So, all machines have a minimum of two NICs? >> > >some machines have 4 nics.. all going to the same nortel switch >( same 1u box ) and connected via is stackable cable in the rear > >some are channel bonded .. most are not > Why do this, especially if they are not bonded? >>With your above idea, why not have the backup server on both LANs? >> > >backups of 500GB during the day kills the network .. which it >does when it goes to the $150K 10-tape library thingie > which in itself should be on a separate tape lan > >in my world, backups should be sorta invisible on the network, >and see if the disk latency is tolerable for head seeks Mine don't. :) Our STK L700 with FC/LTO2 is just dandy and not overloaded at all! Then again, we are back to the system being 8CPU with 8G RAM and multiple gigE nonrouted networks. From alvin at Mail.Linux-Consulting.com Wed Jun 1 14:21:47 2005 From: alvin at Mail.Linux-Consulting.com (Alvin Oga) Date: Wed, 1 Jun 2005 14:21:47 -0700 (PDT) Subject: network thruput In-Reply-To: <429DD113.6090109@halligan.org> Message-ID: hi ya michael On Wed, 1 Jun 2005, Michael T. Halligan wrote: > One quick suggestion. quadruple your # of switches. I've built beowulf > clusters where each server has changing hw is not an option .. esp since reconfig the network or switches is not occuring either > 4-8 nics, each connected to a different switch, and then bonded, > creating a 400 or 800mb (in your case > 4gb or 8gb) network. yes.. the suggestion was to channel bond to go faster.. but its not much of a improvement, since not all machines are bonded > If you have the spare hardware, no spares ... they run 24x7 environment .. no test machines, no hotswap live backup servers either .. yikes > When you say they're getting a paltry 5-10MB/s at best, are you saying > all of the servers are at the > same time, or at any given time? that's the average thruput at any random given time between the supposedly high performance cluster nodes - measured with say copying 100MB or 500MB files between any random node at any random time - to get rid of disk latency issues, we used node1:/dev/loop copying into nodexx:/dev/loop and its the same ... which means the ultra-360 disks is fast enough to keep up on the gigE lan > Beyond that, the stacked switch setup could be bad if that means switch > 10 has to traverse all of the > other switches in order to get to switch 9. that;s the stack i am tryingt break up ... to get rid of all the netbios packets from the cluster ( there is nothing a windoze box needs to do on the cluster ) - netbios packets are about 90% of all packets on the wire > Another thing I'd do is collect some good stats to show to the PHB's .. > Setup NTOP for a week and > show them that it's windows chatter eating up all the bandwidth. If already showed the traffic pattern ... but to no avail ... :-) hard to convince PhD with managerial authority that they're not quite up to puff with network design and topology issues - push too hard, and one is on the streets ya know > they're manageable switches, > setup cacti to graph them via snmp. cacti seems too complicated for me ... :-) i like something simple like ... to show what is clogging the network 90% netbios packets 5% tcpip ( data ) not dns, arp, http, smtp, etc.. 5% misc > Might also be worth digging in to > see if you're having any > type of arp or broadcast storms, perhaps a screwed up vlan. i was hoping to see dns/arp issues but thats not the case here .. > $150k? Ouch. they're very proud to own that $150K tape library... that i will not touch ... not even for $500/hr... no way ... tapes are a disaster waitng to happen in my book and i rather not be restoring from tape or making tape backups, and besides, they have another to take care of that for them > For $20k nowadays you can get a 40 tape lto2 library that > has 200GB (uncompressed) we're looking at 3TB of data .. still pretty small systems actually > I'm starting to give up on Tape to be honest. :-) congrats .. :-) i think after one or few full restores from tapes that someone else did, i think one will no longer be "tape happy" and prefer a more reliable way to restore from full backups ( bare metal restore ) where you have to restore in 5 seconds because the whole company is shutdown until it is back up and online ... - i will always prefer to have live warm-swap backups systems even if i have to bring in my own 2GB - 5GB of disks for those that are willing to pay my fees w/o discounts > The value of tape and disk always goes back and fourth, yes... depending on the sitation fun stuff... c ya alvin From greg.edwards at lmco.com Wed Jun 1 17:02:31 2005 From: greg.edwards at lmco.com (Edwards, Greg) Date: Wed, 01 Jun 2005 17:02:31 -0700 Subject: network thruput Message-ID: <982A2933712F3740921D842654ED470D0DCA3447@emss01m12.us.lmco.com> This discussion would be a wonderful BayLISA presentation, imho. I hope to hear you folks give a talk on this soon. Thank you Greg Edwards -----Original Message----- From: owner-baylisa at baylisa.org [mailto:owner-baylisa at baylisa.org] On Behalf Of Alvin Oga Sent: Wednesday, June 01, 2005 2:22 PM To: baylisa at baylisa.org Subject: Re: network thruput hi ya michael On Wed, 1 Jun 2005, Michael T. Halligan wrote: > One quick suggestion. quadruple your # of switches. I've built beowulf > clusters where each server has changing hw is not an option .. esp since reconfig the network or switches is not occuring either > 4-8 nics, each connected to a different switch, and then bonded, > creating a 400 or 800mb (in your case > 4gb or 8gb) network. yes.. the suggestion was to channel bond to go faster.. but its not much of a improvement, since not all machines are bonded > If you have the spare hardware, no spares ... they run 24x7 environment .. no test machines, no hotswap live backup servers either .. yikes > When you say they're getting a paltry 5-10MB/s at best, are you saying > all of the servers are at the > same time, or at any given time? that's the average thruput at any random given time between the supposedly high performance cluster nodes - measured with say copying 100MB or 500MB files between any random node at any random time - to get rid of disk latency issues, we used node1:/dev/loop copying into nodexx:/dev/loop and its the same ... which means the ultra-360 disks is fast enough to keep up on the gigE lan > Beyond that, the stacked switch setup could be bad if that means switch > 10 has to traverse all of the > other switches in order to get to switch 9. that;s the stack i am tryingt break up ... to get rid of all the netbios packets from the cluster ( there is nothing a windoze box needs to do on the cluster ) - netbios packets are about 90% of all packets on the wire > Another thing I'd do is collect some good stats to show to the PHB's .. > Setup NTOP for a week and > show them that it's windows chatter eating up all the bandwidth. If already showed the traffic pattern ... but to no avail ... :-) hard to convince PhD with managerial authority that they're not quite up to puff with network design and topology issues - push too hard, and one is on the streets ya know > they're manageable switches, > setup cacti to graph them via snmp. cacti seems too complicated for me ... :-) i like something simple like ... to show what is clogging the network 90% netbios packets 5% tcpip ( data ) not dns, arp, http, smtp, etc.. 5% misc > Might also be worth digging in to > see if you're having any > type of arp or broadcast storms, perhaps a screwed up vlan. i was hoping to see dns/arp issues but thats not the case here .. > $150k? Ouch. they're very proud to own that $150K tape library... that i will not touch ... not even for $500/hr... no way ... tapes are a disaster waitng to happen in my book and i rather not be restoring from tape or making tape backups, and besides, they have another to take care of that for them > For $20k nowadays you can get a 40 tape lto2 library that > has 200GB (uncompressed) we're looking at 3TB of data .. still pretty small systems actually > I'm starting to give up on Tape to be honest. :-) congrats .. :-) i think after one or few full restores from tapes that someone else did, i think one will no longer be "tape happy" and prefer a more reliable way to restore from full backups ( bare metal restore ) where you have to restore in 5 seconds because the whole company is shutdown until it is back up and online ... - i will always prefer to have live warm-swap backups systems even if i have to bring in my own 2GB - 5GB of disks for those that are willing to pay my fees w/o discounts > The value of tape and disk always goes back and fourth, yes... depending on the sitation fun stuff... c ya alvin From michael at halligan.org Wed Jun 1 19:01:52 2005 From: michael at halligan.org (Michael T. Halligan) Date: Wed, 01 Jun 2005 19:01:52 -0700 Subject: network thruput In-Reply-To: References: Message-ID: <429E6890.2090501@halligan.org> >>One quick suggestion. quadruple your # of switches. I've built beowulf >>clusters where each server has >> >> > >changing hw is not an option .. > esp since reconfig the network or switches is not occuring > either > > > >>4-8 nics, each connected to a different switch, and then bonded, >>creating a 400 or 800mb (in your case >>4gb or 8gb) network. >> >> > >yes.. the suggestion was to channel bond to go faster.. but >its not much of a improvement, since not all machines >are bonded > > That's a shame, bonding is fun! >> If you have the spare hardware, >> >> > >no spares ... they run 24x7 environment .. no test machines, no >hotswap live backup servers either .. yikes > > > Sigh. Companies never learn. Hardware is so cheap compared to downtime or man hours. >>When you say they're getting a paltry 5-10MB/s at best, are you saying >>all of the servers are at the >>same time, or at any given time? >> >> > >that's the average thruput at any random given time between the >supposedly high performance cluster nodes > - measured with say copying 100MB or 500MB files between > any random node at any random time > > Strange. 10MB/s seems really slow, even with slow hardware and bad disks. My thoughts might be skewed, though, for the past year everything I've played with has been brand new, largely overbuilt, and very fast. > - to get rid of disk latency issues, we used node1:/dev/loop > copying into nodexx:/dev/loop and its the same ... which > means the ultra-360 disks is fast enough to keep up on the > gigE lan > > > >>Beyond that, the stacked switch setup could be bad if that means switch >>10 has to traverse all of the >>other switches in order to get to switch 9. >> >> > >that;s the stack i am tryingt break up ... to get rid of all the netbios >packets from the cluster ( there is nothing a windoze box needs to do >on the cluster ) > > But does NetBios really create that much chatteR? > - netbios packets are about 90% of all packets on the wire > > > >>Another thing I'd do is collect some good stats to show to the PHB's .. >>Setup NTOP for a week and >>show them that it's windows chatter eating up all the bandwidth. If >> >> > >already showed the traffic pattern ... but to no avail ... :-) > >hard to convince PhD with managerial authority that they're not >quite up to puff with network design and topology issues > - push too hard, and one is on the streets ya know > > > >>they're manageable switches, >>setup cacti to graph them via snmp. >> >> > >cacti seems too complicated for me ... :-) > >i like something simple like ... to show what is clogging the network > > 90% netbios packets > 5% tcpip ( data ) not dns, arp, http, smtp, etc.. > 5% misc > > > Cacti is terribly simple. You install the software, point it at your switches, enter in the snmp information, setup devices, setup graphs for those devices' interfaces, and then add them to your device tree. Add a cronjob to poll every 5 minutes, and you've got graphs! >>Might also be worth digging in to >>see if you're having any >>type of arp or broadcast storms, perhaps a screwed up vlan. >> >> > >i was hoping to see dns/arp issues but thats not the case here .. > >>$150k? Ouch. >> >> > >they're very proud to own that $150K tape library... >that i will not touch ... not even for $500/hr... no way ... > >tapes are a disaster waitng to happen in my book and i rather >not be restoring from tape or making tape backups, and besides, >they have another to take care of that for them > > > See. I prefer tape if I'm doing a full-restore and if the full restore is from only one tape.. streaming like that tends to be a lot faster, in HUGE amounts of sequential data, than hard drives, because of seek performance. Unfortunately, it' s more like "go get this 20k file that's 90% into your 200GB tape" .. Ugh. >>For $20k nowadays you can get a 40 tape lto2 library that >>has 200GB (uncompressed) >> >> > >we're looking at 3TB of data .. still pretty small systems actually > > > 3TB of data in one backup? That's annoying. I hate having to backup any individual system that's more than the size of my largest tape. >>I'm starting to give up on Tape to be honest. >> >> > >:-) congrats .. :-) > >i think after one or few full restores from tapes that someone >else did, i think one will no longer be "tape happy" and prefer >a more reliable way to restore from full backups ( bare metal restore ) >where you have to restore in 5 seconds because the whole company >is shutdown until it is back up and online ... > - i will always prefer to have live warm-swap backups systems > even if i have to bring in my own 2GB - 5GB of disks > for those that are willing to pay my fees w/o discounts > > > >>The value of tape and disk always goes back and fourth, >> >> > >yes... depending on the sitation > > >fun stuff... > >c ya > What I find more annoying about tapes, is that tape libraries suck. They're finicky. Cables go bad, their moving parts break, tapes get stuck.. The only thing worse is the terrible complexity of backup software. Every time I have to setup or work with netbackup, I remember just how much I hate veritas' very existance. ------------------- BitPusher, LLC http://www.bitpusher.com/ 1.888.9PUSHER (415) 724.7998 - Mobile From rjholland at ciber.com Wed Jun 1 19:47:07 2005 From: rjholland at ciber.com (Rich Holland) Date: Wed, 1 Jun 2005 22:47:07 -0400 Subject: network thruput In-Reply-To: <429DD113.6090109@halligan.org> Message-ID: <20050602024704.407E38B@orb.sasl.smtp.pobox.com> michael at halligan.org wrote: > $150k? Ouch. For $20k nowadays you can get a 40 tape lto2 library > that has 200GB (uncompressed) tapes.. Another $5-$7k you can go > with the latest SDLT that's 300GB uncompressed (which I'm > looking at purchasing this week, hence the numbers in my head) Where do you guys generally go to get ball-park quotes on tape units? I'm trying to price out several backup options (buy a crateload more DLT-8000 tapes for the 6-yr old library vs. buy a new lto library vs. backup to JBOD). I hate to go through the sales doids and all the dancing just to get ball-park quotes that my customer will likely re-quote through preferred vendors, etc.... Is there anyone on the web advertising pricing (even list price) that I can scrape together something high-level with? Thanks! Rich From michael at halligan.org Wed Jun 1 22:55:24 2005 From: michael at halligan.org (Michael T. Halligan) Date: Wed, 01 Jun 2005 22:55:24 -0700 Subject: network thruput In-Reply-To: <20050602024704.407E38B@orb.sasl.smtp.pobox.com> References: <20050602024704.407E38B@orb.sasl.smtp.pobox.com> Message-ID: <429E9F4C.6000505@halligan.org> Rich, One good way is to look on CDW. When I'm doing a rough estimate on a proposal's bill of sales, I'll go there and see what prices look like. You'd be surprised what CDW actually gives prices for. A lot of enterprise geear is usually unpriceable without picking up the phone, or at least sending off an e-mail. Just like all first prices, the prices on cdw's websites are not the prices you should be paying, if you talk to a sales agent, you'll get better pricing. They're usually able to drop 10-20%.. More on cisco gear which I don't think they're allowed to advertise their real prices. VARS can be your best friend in this situation. I really like doing business with FusionStorm (formerly Cat Technologies). I've been buying them for years now, and they've always treated me well. If I want to compare 3 or 4 products I'll usually send my salesman an e-mail, and get a quote the next day comparing the different options. Over the past year, I've actually found one thing that I like Dell for. They do an OK job of supporting their re-branded ADIC tape libraries. Their prices are phenomenal. I think I paid $13k to buy a dual-drive, 24 slot, lto2 tape unit. Dell's storage support, for tape units, is actually pretty good (I refuse to comment on the rest of their support though.. Life is too short). Their unit doesn't yet support the 300gb sdlt tapes but I'm hoping it will soon, since I need to buy one of these units for a managed contract my company is working on. All in all, CDW is a great place to start. You will find some things that prices don't appear for, but they're really good about getting back to you, without too much fuss, that day with pricing. Michael >michael at halligan.org wrote: > > > >>$150k? Ouch. For $20k nowadays you can get a 40 tape lto2 library >>that has 200GB (uncompressed) tapes.. Another $5-$7k you can go >>with the latest SDLT that's 300GB uncompressed (which I'm >>looking at purchasing this week, hence the numbers in my head) >> >> > >Where do you guys generally go to get ball-park quotes on tape units? I'm >trying to price out several backup options (buy a crateload more DLT-8000 >tapes for the 6-yr old library vs. buy a new lto library vs. backup to >JBOD). > >I hate to go through the sales doids and all the dancing just to get >ball-park quotes that my customer will likely re-quote through preferred >vendors, etc.... > >Is there anyone on the web advertising pricing (even list price) that I can >scrape together something high-level with? > >Thanks! >Rich > > > -- ------------------- BitPusher, LLC http://www.bitpusher.com/ 1.888.9PUSHER (415) 724.7998 - Mobile From alvin at Mail.Linux-Consulting.com Thu Jun 2 01:17:01 2005 From: alvin at Mail.Linux-Consulting.com (Alvin Oga) Date: Thu, 2 Jun 2005 01:17:01 -0700 (PDT) Subject: price quoting - RE: network thruput In-Reply-To: <20050602024704.407E38B@orb.sasl.smtp.pobox.com> Message-ID: hi ya On Wed, 1 Jun 2005, Rich Holland wrote: > michael at halligan.org wrote: > > > $150k? Ouch. For $20k nowadays you can get a 40 tape lto2 library > > that has 200GB (uncompressed) tapes.. Another $5-$7k you can go > > with the latest SDLT that's 300GB uncompressed (which I'm > > looking at purchasing this week, hence the numbers in my head) > > Where do you guys generally go to get ball-park quotes on tape units? I'm > trying to price out several backup options (buy a crateload more DLT-8000 > tapes for the 6-yr old library vs. buy a new lto library vs. backup to > JBOD). i'd use jbod ... if its my choice ... regardless of the size of data for live 24x7 backups for 5 minute downtime sorta thing to come back online AFTER some manual analysis of WHY the primary went down for quotes .. i call 2 or 3 distributor and ask how much, where the "distributor" == "reseller permit required" to buy from them otherwise - generic google/yahoo searches and click on the advertisers to get rough idea .. and web prices tend to be 10-20% below their real prices > I hate to go through the sales doids and all the dancing just to get > ball-park quotes that my customer will likely re-quote through preferred > vendors, etc.... if they are price shopping, i tell them go get their quotes, and i will either match it or pass usually, you will need a dell and hp quote too if you wanna deal with the purchasing dept droids that will be cutting the PO, which they can do themself to get those quotes from their fav var/vad they been buying from for the last 10-20 years > Is there anyone on the web advertising pricing (even list price) that I can > scrape together something high-level with? google ads are good price variation list from vendor to vendor c ya alvin From sigje at sigje.org Thu Jun 2 12:19:14 2005 From: sigje at sigje.org (Jennifer Davis) Date: Thu, 2 Jun 2005 12:19:14 -0700 (PDT) Subject: June 16 2005: DTrace, Bryan Cantrill (BayLISA Monthly) Message-ID: Date: Thursday, 16 June 2005 Where: Apple Computer, Upstairs meeting room in Building 4 (Known as Garage 1) http://www.baylisa.org/locations/current.shtml 7:30 pm Introductions and announcements 8:00 pm Formal presentation 9:45 pm After-meeting dinner/social outing (BJ's, next door) Bryan Cantrill, of Sun will present on DTrace. Full abstract online: http://www.baylisa.org _______________________________________________ baylisa mailing list: baylisa at baylisa.org rsvp for meeting: rsvp at baylisa.org baylisa board (request to sponsor or present): blw at baylisa.org From david at catwhisker.org Tue Jun 7 08:13:50 2005 From: david at catwhisker.org (David Wolfskill) Date: Tue, 7 Jun 2005 08:13:50 -0700 Subject: Requirement for redundant Web service Message-ID: <20050607151349.GM800@bunrab.catwhisker.org> [This is an abbreviated version of a note I just sent to sage-members.] A potential customer (of my employer, an ISP) wants to have an externally-accessible Web server in-house, but also wants requests redirected in case of an outage. I suggested (to the sales person involved on our side) that it was my understanding that Akamia had an offering that looked as if it addressed that requirement. A bit of poking around leaves me with the impression that their "EdgeSuite Delivery" service offering would likely be appropriate. Anyone know of any other services or approaches that would compete with this for the stated requirement? I'll summarize if there's interest expressed. Thanks! Peace, david -- David H. Wolfskill david at catwhisker.org Any given sequence of letters is a misspelling of a great many English words. See http://www.catwhisker.org/~david/publickey.gpg for public key. From david.carmean at netapp.com Tue Jun 7 11:35:31 2005 From: david.carmean at netapp.com (David Carmean) Date: Tue, 7 Jun 2005 11:35:31 -0700 Subject: Recommendation for small personal colo? Message-ID: <20050607183531.GC66080@netapp.com> I'm not excited by the firewall options offered by the virtual server providers I've seen so far (either the packet filtering isn't even stateful, or they're not offering rule customization), so now I'm looking for somewhere between 1 and 4-U space, and since it's just personal and maybe tiny-business use, I don't really need a super-duper Enterprise-grade datacenter. South or East Bay preferred for ease of access. IPv6 connectivity a significant plus. Serial console server with ssh access a major plus for ip_filter maintenance :) Thanks (again)! -- David Carmean Network Appliance, Inc From jxh at jxh.com Tue Jun 7 12:16:35 2005 From: jxh at jxh.com (Jim Hickstein) Date: Tue, 07 Jun 2005 14:16:35 -0500 Subject: Recommendation for small personal colo? In-Reply-To: <20050607183531.GC66080@netapp.com> References: <20050607183531.GC66080@netapp.com> Message-ID: <2A7ACF687D57047A437A3FA9@[192.168.1.100]> > South or East Bay preferred for ease of access. IPv6 connectivity > a significant plus. Serial console server with ssh access a major > plus for ip_filter maintenance :) Look at meer.net. BayLISA is located with them. Besides being competitive on price, they're enormously competent and hence flexible. They've done things for me in the past that other providers would insist were impossible. I used to say, "Announce my route, forward my packets, and get the hell out of my way." Meer does add value (traffic shaping, etc.) but it doesn't feel like they're in my way. From tony at usenix.org Tue Jun 7 12:18:41 2005 From: tony at usenix.org (Tony Del Porto) Date: Tue, 7 Jun 2005 12:18:41 -0700 Subject: Recommendation for small personal colo? In-Reply-To: <20050607183531.GC66080@netapp.com> References: <20050607183531.GC66080@netapp.com> Message-ID: <7dc1ae59c4239747d0f76e1410198ef1@usenix.org> On Jun 7, 2005, at 11:35 AM, David Carmean wrote: > > I'm not excited by the firewall options offered by the virtual > server providers I've seen so far (either the packet filtering isn't > even stateful, or they're not offering rule customization), so now > I'm looking for somewhere between 1 and 4-U space, and since it's > just personal and maybe tiny-business use, I don't really need a > super-duper Enterprise-grade datacenter. > > South or East Bay preferred for ease of access. IPv6 connectivity > a significant plus. Serial console server with ssh access a major > plus for ip_filter maintenance :) > > Thanks (again)! > > > -- > David Carmean Network Appliance, Inc Try: http://www.vix.com/personalcolo/ for a good collection of info on what you are looking for. HTH, Tony From jxh at jxh.com Tue Jun 14 14:27:46 2005 From: jxh at jxh.com (Jim Hickstein) Date: Tue, 14 Jun 2005 16:27:46 -0500 Subject: Kerberos anyone? Message-ID: <4DB9C8836A0BC0B086B7C1AB@[10.9.18.3]> Who knows anything about Kerberos? I want a consultant to set up a Kerb4/5 server in a hurry for testing something. Bay area, one day. Please email me privately. Thanks! From sigje at sigje.org Thu Jun 16 11:07:03 2005 From: sigje at sigje.org (Jennifer Davis) Date: Thu, 16 Jun 2005 11:07:03 -0700 (PDT) Subject: June 16 2005: DTrace, Bryan Cantrill (BayLISA Monthly) In-Reply-To: References: Message-ID: Don't forget!!! Tonight is BayLISA! Date: Thursday, 16 June 2005 Where: Apple Campus, Upstairs meeting room in Building 4 (Known as Garage 1) http://www.baylisa.org/locations/current.shtml 7:30 pm Introductions and announcements 8:00 pm Formal presentation 9:45 pm After-meeting dinner/social outing (BJ's, next door) Bryan Cantrill, of Sun will present on DTrace. Full abstract online: http://www.baylisa.org From jxh at jxh.com Fri Jun 17 14:28:53 2005 From: jxh at jxh.com (Jim Hickstein) Date: Fri, 17 Jun 2005 16:28:53 -0500 Subject: BayLISA webcal? Message-ID: <21BE26AC661C493B2D9FD098@[10.9.18.3]> I just fired up iCal for the first time[1]. Apparently one can subscribe to published calendars -- things like US Holidays, or events at your nearby Apple Store. :-) We should have the BayLISA calendar up! How does one go about it? -- [1] I know, I'm far behind on this stuff. But I sleep like a log knowing my address book isn't in the clutches of the likes of Time Warner or NSI. I'm deeply suspicious even of using .Mac just to shove things to another computer 3 feet away. Screw that. From bill at wards.net Fri Jun 17 16:03:38 2005 From: bill at wards.net (William R Ward) Date: Fri, 17 Jun 2005 16:03:38 -0700 Subject: Recommendation for FreeBSD virtual private server vendor? In-Reply-To: <20050601164816.GO27610@netapp.com> References: <20050601164816.GO27610@netapp.com> Message-ID: <17075.22218.747986.747687@komodo.home.wards.net> David Carmean writes: > >Feeling guilty about the spam and diskspace load I'm putting on my friends' >home/DSL server, I'm looking into affordable alternatives and Virtual Private >Servers are looking interesting. I prefer FreeBSD over Linux. > >What I'm concerned about is how to get one set up with packet filters (pref. >ip_filter) before the thing goes on-line. How do vendors handle that >chicken-and-egg problem? > >Any recommendations for vendors or what to look for in services? RackSpace's >DDS looks interesting. > >Is anybody reselling VMWare-based virtual machines? I know you said you prefer not to use Linux, but I have no such qualm and am fairly happy with linode.com. As for packet filters, the default installation of most modern OS's only opens up a handful of key ports anyway, such as ssh and http, so it shouldn't be a big risk. Plus there's nothing there yet for them to hack and you know it's the most current version of the OS. -- William R Ward bill at wards.net http://bill.wards.net ----------------------------------------------------------------------------- Help save the San Jose Earthquakes - http://www.soccersiliconvalley.com/ From michael at halligan.org Fri Jun 17 18:56:47 2005 From: michael at halligan.org (Michael T. Halligan) Date: Fri, 17 Jun 2005 18:56:47 -0700 Subject: Recommendation for FreeBSD virtual private server vendor? In-Reply-To: <17075.22218.747986.747687@komodo.home.wards.net> References: <20050601164816.GO27610@netapp.com> <17075.22218.747986.747687@komodo.home.wards.net> Message-ID: <42B37F5F.5050608@halligan.org> You should consider www.hub.org .. It's a pretty-well run site which uses FreeBSD (I believe0, and part of it's money goes to help the postgres project. Another good choice is Aktiom, but I'm not sure if they provide FreeBSD servers or not. William R Ward wrote: >David Carmean writes: > > >>Feeling guilty about the spam and diskspace load I'm putting on my friends' >>home/DSL server, I'm looking into affordable alternatives and Virtual Private >>Servers are looking interesting. I prefer FreeBSD over Linux. >> >>What I'm concerned about is how to get one set up with packet filters (pref. >>ip_filter) before the thing goes on-line. How do vendors handle that >>chicken-and-egg problem? >> >>Any recommendations for vendors or what to look for in services? RackSpace's >>DDS looks interesting. >> >>Is anybody reselling VMWare-based virtual machines? >> >> > >I know you said you prefer not to use Linux, but I have no such qualm >and am fairly happy with linode.com. > >As for packet filters, the default installation of most modern OS's >only opens up a handful of key ports anyway, such as ssh and http, so >it shouldn't be a big risk. Plus there's nothing there yet for them >to hack and you know it's the most current version of the OS. > > > -- ------------------- BitPusher, LLC http://www.bitpusher.com/ 1.888.9PUSHER (415) 724.7998 - Mobile From michael at halligan.org Sat Jun 18 09:22:33 2005 From: michael at halligan.org (Michael T. Halligan) Date: Sat, 18 Jun 2005 09:22:33 -0700 Subject: SIG-BEER-WEST? In-Reply-To: <20041211230623.A13661@gerasimov.net> References: <20041211230623.A13661@gerasimov.net> Message-ID: <42B44A49.8020906@halligan.org> I'm curious, whatever happened to sig-beer-west? Has this fizzled out? If so, I say let's revive it tonight at the Crow Bar in North Beach. Michael From michael at halligan.org Sat Jun 18 09:31:27 2005 From: michael at halligan.org (Michael T. Halligan) Date: Sat, 18 Jun 2005 09:31:27 -0700 Subject: Anybody interested in going in on some bandwidth from Internap? Message-ID: <42B44C5F.5080309@halligan.org> Greetings, I provide personal colo @365main.net to a number of local sysadmins. Right now, we're using netoptex as our main bandwidth, but I have a need to get a few mb/s from Internap for a project I'm working on. I was wondering if anybody is in the similar situation, or would be interested in throwing in a server at a good datacenter and buying some at-cost internap bandwidth. I'm trying to get my commit up to a "reasonable" level so we can get a good price break. If anybody is interested, let me know. Michael ------------------- BitPusher, LLC http://www.bitpusher.com/ 1.888.9PUSHER (415) 724.7998 - Mobile From cheselka at gmail.com Sat Jun 18 18:15:02 2005 From: cheselka at gmail.com (Michael Cheselka) Date: Sat, 18 Jun 2005 18:15:02 -0700 Subject: SIG-BEER-WEST? In-Reply-To: <42B44A49.8020906@halligan.org> References: <20041211230623.A13661@gerasimov.net> <42B44A49.8020906@halligan.org> Message-ID: <3c6fb88105061818155b267343@mail.gmail.com> Hello, Mssr. David Alban has a job now and therefore less time. Anyone care to pick it up, or organise a round-robin so no one person has the responsiblitiy? Mike On 6/18/05, Michael T. Halligan wrote: > I'm curious, whatever happened to sig-beer-west? Has this fizzled out? > > If so, I say let's revive it tonight at the Crow Bar in North Beach. > > Michael > -- When in the Potemkin Wired be a Potemkin Lain. From extasia at gmail.com Sat Jun 18 19:53:16 2005 From: extasia at gmail.com (David Alban) Date: Sat, 18 Jun 2005 19:53:16 -0700 Subject: kill -ROUND-TUIT $SBW_PID [was: SIG-BEER-WEST?] In-Reply-To: <42B44A49.8020906@halligan.org> References: <20041211230623.A13661@gerasimov.net> <42B44A49.8020906@halligan.org> Message-ID: <4c714a9c05061819533d91bb43@mail.gmail.com> The Instigator is glad you asked... sbw was starved of cycles when, over the course of several months, one of The Instigator's family members was taken ill and passed on. The Instigator promised The Instigator that The Instigator would pick it (sbw) up again after taking a couple months off. Long past that point now, The Instigator has experienced difficulties finding a Round TUIT, which is why The Instigator is glad you asked. sbw's pilot light indeed went out, but The Intent was always to relight it. The Instigator is duly (and cheerfully) admonished and promises to coordinate a July event. sbw will be moving mostly likely to 2nd Saturdays, as The Instigator has a standing commitment on 3rd Saturdays. The Instigator asks that any respondents to MTH's email that find themselves in the Crow Bar tonight lift a glass and toast to Those Mysterious Occurrences which connect people with their Round TUITs. On 6/18/05, Michael T. Halligan wrote: > I'm curious, whatever happened to sig-beer-west? Has this fizzled out? > > If so, I say let's revive it tonight at the Crow Bar in North Beach. -- Live in a world of your own, but always welcome visitors. From extasia at gmail.com Sat Jun 18 20:03:08 2005 From: extasia at gmail.com (David Alban) Date: Sat, 18 Jun 2005 20:03:08 -0700 Subject: SIG-BEER-WEST? In-Reply-To: <3c6fb88105061818155b267343@mail.gmail.com> References: <20041211230623.A13661@gerasimov.net> <42B44A49.8020906@halligan.org> <3c6fb88105061818155b267343@mail.gmail.com> Message-ID: <4c714a9c05061820034faaeb72@mail.gmail.com> On 6/18/05, Michael Cheselka wrote: > Mssr. David Alban has a job now and > therefore less time. Anyone care to pick it up, or organise a round-robin > so no one person has the responsiblitiy? The Instigator has in the past occasionally needed to solicit the services of Guest Hosts when the Instigator could not be present at an event. Anyone wishing occasionally to be a Guest Host should feel free to inform The Instigator of that fact. The Instigator is *always* warmly receptive to suggestions for venues, requirements for which being: * reasonable Bay area location * superior selection of beer (very, very important) The Instigator is always open to discussing changes in the sbw process. However, such changes have to be approved by the sbw board, whose sole member is The Instigator. -- Live in a world of your own, but always welcome visitors. From bill at thecrookes.com Tue Jun 21 08:42:18 2005 From: bill at thecrookes.com (Bill Crooke) Date: Tue, 21 Jun 2005 08:42:18 -0700 Subject: Peninsula Linux Users' Group Meeting, Thursday, June 23rd, 2005 Message-ID: <42B8355A.8010108@thecrookes.com> Peninsula Linux Users' Group, Thursday, June 23, 2005 We have a meeting of the Peninsula Linux Users' Group (PenLUG) this week! Here are the details about the next meeting. For more information or directions go to http://www.penlug.org/ Our website is a TWiki; please feel free to create a user account and modify the website if you have something to contribute. Date: Thursday, June 23rd, 2005 Time: 7:00 - 9:00 PM Location: 100 Oracle Parkway, Redwood Shores, CA 94065 Room 104 Agenda: ======= 7:00 - 8:30 PM: Presentation by Bart Decrem: "Firefox and the Linux Desktop" 8:30 - 9:00 PM: Members' Minutes 8:45 - 9:00 PM: Adjourn to IHOP (Belmont) for social & food time Presentation by Bart Decrem: "Firefox and the Linux Desktop" ====================================================== Bart Decrem has contributed tremdously to Linux on the Desktop through his work at companies such as Eazel and Hancom Linux, as well as through his work with the GNOME and Mozilla Foundations. At the Mozilla Foundation Bart oversaw marketing and business affairs for the Foundation, including the "Spread Firefox" campaign which has been instrumental in the rapid growth of the Firefox browser. In April Bart started a new company called Round Two whose mission is to provide "a new crop of products and services that will enhance your Firefox experience". Bart's talk will cover the rapid growth of Firefox, it's implications for the growth of Desktop Linux, and Round Two's plans for enhancing the Firefox experience. Members' Minutes ================ Members will have an opportunity to take a few minutes to... * Describe their latest Linux discovery * Ask questions and get help from other members * Discuss Linux projects You can just stand up and talk, or give a short demo or presentation. If you need audio/visual support for your Members' Minute, please contact me in advance to arrange for your needs. We have a limited number of books courtesy of Prentice-Hall and O'Reilly to give away as an added inducement to participate in this portion of the meeting. :-) 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. Bill Crooke PenLUG Speaker Coordinator From michael at halligan.org Tue Jun 21 16:26:23 2005 From: michael at halligan.org (Michael T. Halligan) Date: Tue, 21 Jun 2005 16:26:23 -0700 Subject: Distributed/Clustered file systems? Message-ID: <42B8A21F.1080401@halligan.org> I'm looking for some low-cost alternatives to buying a SAN/NAS, or even a large file server. I'm about to do a tech refresh, buying a dozen new servers and relinquishing our old servers to our testing network. I'm probably going to buy an exabyte lto-3 tape library, but instead of the complexity of host-based backup systems, I'd like to use the free space on each server (of which there will be quite a bit) to create a couple of large volumes, effectively doing RAID over my lan. The only options that seem to be viable for this kind of setup are PVFS and Lustre. Does anybody have experience with this type of solution, and can recommend one from a stability point of view? ------------------- BitPusher, LLC http://www.bitpusher.com/ 1.888.9PUSHER (415) 724.7998 - Mobile From alvin at Mail.Linux-Consulting.com Wed Jun 22 01:08:14 2005 From: alvin at Mail.Linux-Consulting.com (Alvin Oga) Date: Wed, 22 Jun 2005 01:08:14 -0700 (PDT) Subject: Distributed/Clustered file systems? In-Reply-To: <42B8A21F.1080401@halligan.org> Message-ID: On Tue, 21 Jun 2005, Michael T. Halligan wrote: > I'm looking for some low-cost alternatives to buying a SAN/NAS, or even > a large file server. I'm about to do > a tech refresh, buying a dozen new servers and relinquishing our old > servers to our testing network. I'm probably > going to buy an exabyte lto-3 tape library, but instead of the > complexity of host-based backup systems, I'd like > to use the free space on each server (of which there will be quite a > bit) to create a couple of large volumes, effectively > doing RAID over my lan. The only options that seem to be viable for > this kind of setup are PVFS and Lustre. or network raid ( drbd.org ) or any "global filesystem" or "distributed fs" ?? - which one to use it the "trick" > Does > anybody have experience with this type of solution, and can recommend > one from a stability point of view? too new to be the primary fs of corp data ... good for secondary backups to test against other backups ( 100TB live backups are gonna be fun ) c ya alvin From strata at virtual.net Thu Jun 23 17:40:17 2005 From: strata at virtual.net (Strata R. Chalup) Date: Thu, 23 Jun 2005 17:40:17 -0700 Subject: [Fwd: [SAGE] re: SAGE Voting Site] Message-ID: <42BB5671.9010102@virtual.net> FYI, if you are a current SAGE member, please take a moment to vote for the new board. Voting ends this week, on Friday at 5pm PDT. Details are at http://www.sage-members.org/election/ cheers, Strata -------- Original Message -------- Subject: [SAGE] re: SAGE Voting Site Date: Thu, 23 Jun 2005 16:27:12 -0500 From: David Parter To: sage-members at sage.org Some questions have been raised about the SAGE voting site, and I think I should clarify the situation. It is unfortunate that this information was not presented in advance. It is one of those things that is obvious after the fact, but slipped through in the run-up to the election. First, it is important to understand that the SAGE Interim Board, as candidates (and the other candidates) have taken as much of a hands-off approach to the election logistics as possible, in order to avoid the appearance of any conflict of interest or election improprieties. I also have been reluctant to get involved in most of this discussion while the election is in progress, because it might seem to be an attempt to gain an unfair advantage during the election. However, I've been asked by enough members to explain things, so here it is. My role in this was as the coordinator/contact to the volunteer tech team that was recruited a few months ago to help the Interim Board/Transition Committee deal with IT/technical issues. There were a couple of issues regarding hosting the election at www.sage.org: 1) In the past two SAGE STG online elections, the election was hosted at a "neutral" site -- one not connected to any of the candidates, Board or staff -- for the same reason that the candidates took a hands-off approach: to avoid the appearance of any conflict of interest or election improprieties. We felt it appropriate to continue that approach 2) Asking a volunteer to undertake a substantial and critical technical task, but making them do it on a site that they do not have access to, and are not familiar with, is a recipe for disaster. Any sysadmin who's ever tried to install something on someone else's system should understand the issues. Add to that the known complexity and fragility of the existing SAGE web site, and the need to coordinate with the USENIX staff several times zones away if there were any issues, and it does make sense. Instead, we asked our volunteers (SAGE members Jesse Trucks and Matt Okeson-Harlow of Cyberius' network) to meet certain criteria for a fair, secure election, including assurances about the security of the site, the integrity of the election, and making the source code available for inspection. After recruiting Matt and Jesse, the coordination of the technical aspects of the election site was between Matt, the Leadership Committee, and staff (especially Jane-Ellen Long) in the USENIX office who provided the membership data used for authentication and authorization. The sage-members.org domain was already owned by a SAGE member, who gaciously volunteered to let the Interim Board use it for our interim internet needs, including the Board's wiki, mailing lists, and eventually the candidate's wiki (which was opened to everyone instead of copying the material to another site) and the election information site. The content of www.sage-members.org/election, including the description of the voting method, was left to the Leadership Committee. I'm sure that future boards will apply the lessons learned from the election in organizing future elections. Thank you everyone who has participated in this discussion for their concerns about the organization. If you haven't already, please vote for the candidates of your choice (ranked from 1 to N, of course!), and please pass the word on to other sage members who may have missed the announcements. --david David Parter SAGE Interim Board SAGE Board candidate -- ======================================================================== Strata Rose Chalup [KF6NBZ] strata "@" virtual.net VirtualNet Consulting http://www.virtual.net/ ** Strategic IT for the Growing Enterprise ** ========================================================================= From michael at halligan.org Sat Jun 25 20:32:45 2005 From: michael at halligan.org (Michael T. Halligan) Date: Sat, 25 Jun 2005 20:32:45 -0700 Subject: Published calendars Message-ID: <42BE21DD.1030604@halligan.org> My apologies if you receive this message several times, I've cc:'d several local lug/lisa groups. How about publishing the group meeting & event calendars via ical? It'd be very nice to be able to just subscribe to everry lug/lisa/sage/usenix calendar and be able to schedule my own events with knowledge of group happenings. ------------------- BitPusher, LLC http://www.bitpusher.com/ 1.888.9PUSHER (415) 724.7998 - Mobile From sigje at sigje.org Mon Jun 27 11:14:44 2005 From: sigje at sigje.org (Jennifer Davis) Date: Mon, 27 Jun 2005 11:14:44 -0700 (PDT) Subject: OpenSolaris meeting - Bay area Message-ID: Location: Sun Santa Clara Auditorium (upstairs SCA03) When: Tuesday, June 28th, 2005 (Tomorrow!) Topic: Open Solaris Time: 7:30pm-10:00pm June's BayLISA talk was Bryan Cantrill covering some nice features of DTrace on his newly installed OpenSolaris OS laptop. For people who are interested in learning more about Open Solaris (and live in the San Francisco/Santa Clara area) this might be very insightful! Jennifer