Back to Home / #linode / 2006 / 01 / Prev Day | Next Day
#linode IRC Logs for 2006-01-23

---Logopened Mon Jan 23 00:00:59 2006
00:37|-|dc0e [] has quit [Remote host closed the connection]
00:38|-|dc0e [] has joined #linode
02:25|-|womble [~mpalmer@] has joined #linode
04:13|-|Dreamer3 [] has quit [Ping timeout: 480 seconds]
04:17|-|Dreamer3 [] has joined #linode
04:18|-|womble [~mpalmer@] has quit [Quit: What's behind the round window...]
05:49|-|internat [] has quit [Quit: This computer has gone to sleep]
06:18|-|jekil [] has joined #linode
06:21<linbot>New news from forums: Linode 80 LAMP Capacity Formula Question in Linux, Apache, Mysql and PHP (LAMP) Forum <>
06:27<linbot>New news from forums: lighttpd instead of Apache? in Linux, Apache, Mysql and PHP (LAMP) Forum <>
06:37|-|linville [] has joined #linode
07:42|-|jekil [] has quit [Quit: Leaving]
07:49|-|internat [] has joined #linode
08:28<warewolf>god windows sucks
08:28<warewolf>jesus christ
08:28<warewolf>I only wish this machine running windows had usb 2.0
08:28<warewolf>then I could copy tivo recordings off it fast as hell to my usb hd
08:29|-|Redgore [~Redgore@] has joined #linode
08:34<warewolf>how are the files bigger on the USB HD than on the windows machine
09:00|-|emcnabb [] has quit [Quit: IRC: Where men are men, women are men, and little girls are FBI agents]
09:12<npmr>warewolf, are the devices formatted using different filesystems?
09:12<npmr>warewolf, do the files contains sparse areas?
09:12<npmr>warewolf, are your tools reporting the actual file size or the number of blocks allocated?
09:13<npmr>warewolf, are the block sizes the same
09:22<internat>so if i have a rdiff backup.. how do i get a file out of it that i deleted?
09:42|-|spr [] has joined #linode
09:42<linbot>New news from forums: Modules busted all of a sudden in General Discussion <>
09:43|-|anderiv [] has quit [Quit: leaving]
09:44|-|anderiv [] has joined #linode
09:46|-|FireSlash [] has joined #linode
10:25|-|FireSlash [] has quit [Quit: Leaving]
10:43|-|jekil [~alessandr@] has joined #linode
11:28|-|besonen [] has quit [Quit: Leaving]
11:37|-|besonen [] has joined #linode
11:46|-|Newsome [] has joined #linode
12:13|-|jekil [~alessandr@] has quit [Ping timeout: 480 seconds]
12:18|-|jekil [] has joined #linode
12:50|-|jekil [] has quit [Quit: Leaving]
13:01|-|FireSlash [] has joined #linode
13:36|-|jekil [~alessandr@] has joined #linode
13:38<@caker> <-- video documentary on Google
13:43|-|FireSlash [] has quit [Read error: Connection reset by peer]
14:07|-|Redgore [~Redgore@] has quit [Quit: A geek without purpose - | SMDC-Network IRC -]
14:08|-|Redgore [~Redgore@] has joined #linode
14:08|-|tsi [] has joined #linode
14:09<tsi>dunno if it's been reported, but host42 is explodeying
14:11[~]caker looks
14:12<tsi>i could tell when you started looking, my session went from slow to stop
14:12<@caker>sssh! we're hunting swap thrashers
14:12<tsi>cwush those swap twashews!
14:15<@caker>better now?
14:16<tsi>load dropping from 5
14:16<tsi>load average: 0.08,
14:16<@caker>cool .. thanks for letting me know
14:24|-|FireSlash [] has joined #linode
14:32<[|^__^|]>HOSTY KERSPLODEY
14:32<[|^__^|]>(not really)
15:01<fo0bar>caker: hey, mwave is carrying supermicro opteron servers now
15:02<FireSlash> <-- I want one of these. Fully loaded.
15:03<fo0bar>aka "supermicro is forcing us to not just call them 'supermicro opteron systems' because supermicro is scared of intel pushing them down the stairs"
15:03<Redgore>FireSlash: I want a ISM zSeries
15:04[~]Redgore can imagine having it sat in the corner of his room crunching proteins
15:04<FireSlash>I'm content with 128 processors, 512gb of ram, and a bus speed of 133GB/s
15:07<Redgore>its more the geek factor than the power
15:09[~]tronix would love to run z/OS on nice hardware like the z990... :-)
15:09<tronix>z900, in a maxed out config could theoretically hit about 1/2 million I/Os per sec
15:10<tronix>I've got older hardware (pre-zSeries) @ work. bahh.
15:12<tronix>speaking of opterons... we've got some Sun 1U opteron boxes at work, low end stuff
15:12<tronix>and they run between 30x to 40x faster than the 2U machines they replaced. sweet stuff
15:12<tronix>1/2 the weight, 1/2 the height, 1/3 the price, 30-40x the perf
15:12<tronix>these boxes are bored silly now. :( load avg is about 0.01 or 0.02
15:13<tronix>the busiest box? 0.08
15:13<@caker>fo0bar: finally :)
15:14|-|jekil [~alessandr@] has quit [Ping timeout: 480 seconds]
15:15<@caker>so .. I'll only need another vendor for the 3ware card
15:15<@caker>oh, and memory
15:16<@caker>POLL: Do you find it annoying when people refer to disk storage as "memory"?
15:17<@caker>same here
15:17<anderiv>though I usually hear that from not-so-tech-savvy folks, so I can't fault them too much.
15:17<anderiv>...and I usually try and correct them.
15:19<tronix>caker: heh... the only time I give them a free pass is when they're AS/400 folks :-)
15:20<tronix>(OS/400 has something called SLS -- Single Level Store; it combines physical + swap in interesting ways)
15:20<@caker>aww, NBC is cancelling West Wing :(
15:20<tronix>but for everybody else, that's annoying!
15:21|-|Newsome [] has left #linode [Linux: Now with special employee pricing!]
15:22<Redgore>when RAM drives eventually come public, then storage could be called memory :P
15:27<[|^__^|]>well realize that it used to be that way
15:27<[|^__^|]>like, you had core and you had tape
15:27<[|^__^|]>tape was for backup
15:27<[|^__^|]>core was all the storage you had otherwise
15:28<[|^__^|]>it was a more obvious metaphor in the 60s
15:33<linbot>New news from forums: [ Poll ] Memory != disk space in /dev/random <>
15:51|-|rafa [] has joined #linode
15:52|-|jekil [] has joined #linode
15:53<rafa>hi all
16:00|-|jekil [] has quit [Ping timeout: 480 seconds]
16:03<@mikegrb>mmm cake
16:03<rafa>cake ?
16:03<rafa>mike ?
16:06|-|besonen [] has quit [Ping timeout: 480 seconds]
16:07<fo0bar>caker: you know what the ironic thing is? I ordered a personal server (P4) from mwave about 3 hours before that section appeared :)
16:08<fo0bar>I still probably would have bought the P4. Opterons are kickass value compared to Xeons, but a P4 would still have better performance-to-cost
16:08<fo0bar>what I *really* want to see on the market are athlon64 server boards
16:09|-|besonen [] has joined #linode
16:13<fo0bar>caker: and re: memory vendor... who do you go through and why? I'm slightly ashamed to admit, but I use the "mwave brand memory" (which is to say, whatever they have around at the time)
16:13<fo0bar>usually it's kingston valueram
16:13<fo0bar>but I haven't had a problem yet
16:13[~]fo0bar knocks on wood
16:14<fo0bar>"mwave brand power supplies", on the other hand... 2 out of 2 failures :P
16:15<fo0bar>hint: a PSU should not smell like burning electronics
16:15<fo0bar>and the 12v rails should not be 10.2v
16:15|-|Redgore [~Redgore@] has quit [Quit: A geek without purpose - | SMDC-Network IRC -]
16:25|-|Dreamer3 [] has quit [Ping timeout: 480 seconds]
16:26|-|Dreamer3 [] has joined #linode
16:41|-|rafa [] has quit [Quit: rafa]
16:42|-|J[SS] [] has quit [Read error: Connection reset by peer]
16:44|-|J[SS] [] has joined #linode
16:50|-|FireSlash [] has quit [Ping timeout: 480 seconds]
16:50<fo0bar>100%[==================================================================================>] 712,204,288 7.68M/s ETA 00:00
16:56|-|FireSlash [] has joined #linode
17:03|-|FireSlash [] has quit [Quit: Leaving]
17:04|-|sec39 [] has quit [Ping timeout: 480 seconds]
17:19|-|sec39 [] has joined #linode
17:38|-|cow [] has quit [Ping timeout: 480 seconds]
17:52|-|jekil [] has joined #linode
18:12|-|linville [] has quit [Quit: Leaving]
18:27|-|FireSlash [] has joined #linode
18:29|-|spr [] has quit [Quit: Spoon!]
18:31|-|rko [] has quit [Ping timeout: 481 seconds]
18:34|-|jekil [] has quit [Quit: Leaving]
18:49|-|darkbeholder [] has quit [Read error: Connection reset by peer]
18:56|-|darkbeholder [] has joined #linode
19:29<@caker>fo0bar: did you notice the new A+ boxes?
19:35<@caker>fo0bar: looks like the major difference in the various new boxes is that they're all using ServerWorks chipset (SATA, NIC, etc), and one model has hot-swap redundant PSUs
19:50<tronix>for anybody familiar with redhat (guessing rh8 or rh9?), how do you disable dhcp client and stick in static IPs?
19:50<tronix>I haven't run RH in so long and someone was asking
19:50<tronix>something like /etc/sysconfig/network ?
19:50<@caker>tronix: /etc/sysconfig/network-scripts/ifcfg-eth0
19:50<tronix>ahh that's it. thanks!!
19:50<@caker>tronix: and /etc/sysconfig/network
19:51<tronix>sweet. should be e-z from here on out. thanks!
19:59<@mikegrb>Obtain can of shaving cream (pressurised).
19:59<@mikegrb>Leave can in freezer for about a week.
19:59<@mikegrb>Remove can from frozen contents using tin-snips or similar.
19:59<@mikegrb>Quickly put contents in offending vehicle.
20:01[~]mikegrb should do that in the backyard
20:05<internat>so how hard would it be to setup a script that determins how much hdd space is left on my node, and when it drops below a certian size it does somethingg
20:05<@mikegrb>me suggests rm -rf /usr for the does something part
20:06<internat>generally by does something, it will be a sms notification :)
20:06<@mikegrb>followed by rm -rf / if the situation isn't improved
20:06<internat>saying omfg weve run outa space, hacks hack1!!
20:06[~]mikegrb suggests df for the determines how much hdd space is left part
20:08<internat>man i need to learn sed and regex so i can figure out how to parse shit out
20:08<@mikegrb>awk is probably more of what you want
20:08<internat>err, ok?
20:10<@mikegrb>sed is more for search and replace
20:10<@mikegrb>awk is more of search for something and then do something with it
20:11<fo0bar>caker: I noticed, but I didn't look into them too hard
20:11<internat>ah ok'
20:11<@mikegrb>df / | awk '/\// { print $5 }' <-- prints % usage of /
20:11<@caker>fo0bar: yeah .. no messing with Marvell drivers and whatnot. Depending on how kick-butt this 9550SX card turns out, I might not need to bother with that, either
20:12<@caker>fo0bar: btw, I had to trim the CPU shroud to route the SATA cables to the card .. it was a tight fit
20:12<fo0bar>caker: I was a bit confused by that though... usually supermicro only adds a new line when new CPUs are added (IE, the 5013 -> 5014 line for 533mhz -> 800mhz P4s)
20:13<@caker>fo0bar: so far, I've only figured out that ... AS1020A-T, AS is a+ supermicro, the first 1 is 1U, the 2 is 2-cpu
20:13<fo0bar>caker: interesting. the only thing I have to do with 5014 models are ditch the short cables because they're usually just a millimeter or 2 too short
20:13<fo0bar>which doesn't matter, I have hundreds of sata cables at work
20:14<fo0bar>I should take a picture of the bin full of them
20:14<@caker>fo0bar: yeah, that's just because these AMD boxes have a giant shroud that goes pretty deep back into the case
20:14<fo0bar>half of me says "uhh, you'll never use all those cables", the other have says "no! build your collection!"
20:15<@caker>fo0bar: hah .. I have a collection of those small cables, as well .. along with my heatsink and intel sticker collection
20:16<@caker>no clue what the A-T, A-8, S-8, P-8, P-T, P-TR endings refer to
20:16<fo0bar>yeah. there's no room for stickers on a supermicro 1U box, so I have a collection too. at first I started putting them ironic locations (on a switch, ip phone, mac, xeon sticker on a 300mhz K6)
20:16<fo0bar>but I ran out of irony
20:16<@caker>fo0bar: btw, when did you last look? Cause they added more a few weeks back, and now there's 5 more added in the last few days
20:17<fo0bar>caker: the end usually refers to the drive options and type. like the ones with "8" in them are scsi
20:17<fo0bar>"T" usually refers to sata
20:18<fo0bar>caker: today, after I noticed the mwave offerings
20:19<encode>fo0bar: hehe, we have drawers full of spare scsi cables that we'll never use too
20:20<encode>cos practically everything that needs a scsi cable comes with one already
20:22<fo0bar>encode: our company began an anti-scsi campaign a couple years back, because of a stream of EXTREMELY bad luck with scsi drives. now, whenever a SCSI system is taken out of a rack (even if it's not to completely replace it), the controller and drives are replaced with SATA
20:23<fo0bar>because of this, I have a bigass plastic tub in the back room labeled "Box 'o SCSI"
20:23<fo0bar>dozens of controllers, even more drives, hundreds of cables
20:24<internat>web design sucks wang :/
20:24<encode>our servers are almost exclusively hp, with seagate drives
20:24<encode>we dont have a terribly high failure rate
20:24<fo0bar>I keep thinking about ebaying them, but I would either have to 1) sell as-is and get jack for each thing sold, or 2) verify what stuff is and isn't working, and still possibly get jack when sold
20:25<fo0bar>seagate drives are nice, all of our sata stuff is seagate
20:25<encode>or just send them to me, and i'll pay postage
20:25<encode>and you'll get nothing for them mwahah
20:26<fo0bar>yeah, well, they're company assets, so I'd have to work with finance to do that, and they'd probably want to try to sell them
20:26<encode>yeah i know
20:26<encode>i wouldnt have any use for them personally anyway
20:27<fo0bar>probably conversation: "sell them! even if we don't get much, we'll still get money!" me: "the labor involved in selling them would probably cost more than we'd get"
20:27<encode>just put them in a crate
20:27<internat>i have a question
20:27<encode>and sell them by weight on ebay
20:27<internat>since currently ive never mananged to do it.. what happens when u go over ur network transfer?
20:27<fo0bar>same deal on a stack of about 20 1U 500mhz celeron servers
20:27<internat>do u get blocked? or do u start getting charged?
20:28<fo0bar>internat: you're billed based on the standard bandwidth overage charge
20:28<internat>okies :) i can deal with that
20:28<encode>ooh, 1u 500Mhz celerons
20:28<encode>i could use a couple of them
20:28<fo0bar>internat: but IIRC, if it's a simple mistake and you're nice, they'll forgive you once
20:28<encode>unfortunately, the same labor and finance dept stuff applies
20:29<encode>and shipping to .au would be expensive++
20:29<fo0bar>encode: my boss's boss wants to have a "employee auction" to get rid of them. after that fails (I know of noone that wants one in the office), I'll see if I can do anything :)
20:30<internat>dammit i need to get some rj45 shit so i can wire up my house
20:30<encode>fo0bar: nah, dont worry about it, more hastle that its worth
20:30<fo0bar>hell, I have 2 800mhz P3 1Us sitting in a corner at home, and I just bought a 3ghz 1U for actual use
20:31<fo0bar>BTW, I have a dirty secret: I'll be moving most of my linode stuff to a real colo (but I'll be keeping the linode as well)
20:31<fo0bar>*maybe* drop it down from a $60 plan to a $40 plan, but we'll see
20:31<encode>:O blasphemy!
20:32<fo0bar>but you know what the sweet part is? my linode and the new provider are 4ms away from each other :)
20:32<@mikegrb>internat: df / | awk '/\// { if (substr($5, 1, length($5)-1) < 10) print "Less then 10% free - " $5 }'
20:32<fo0bar>and I can get about 60mbps between them
20:33<internat>im assuming that only does something if its less then 10%
20:33<encode>fo0bar: nice
20:33<fo0bar>internat: awk is mostly deep voodoo for me
20:33<fo0bar>but I would assume so too
20:34<encode> <-- anyone ever done something like this?
20:34<@mikegrb>internat: prints: Less then 10% free - X%
20:34<@mikegrb>otherwise prints nothing
20:36<@mikegrb>you could replace the print blah blah bit at the end with: system("/usr/local/bin/") and have page you
20:37<@mikegrb>likely: system("/usr/local/bin/ '" $5 "'") would execute /usr/local/bin/ 3%
20:37<@mikegrb>or whatever percent of space was free so that it could be included in the message
20:38<@mikegrb>er ' around the 3% like /usr/local/bin/ '3%'
20:38<@mikegrb>but I've never used system in awk
20:39<fo0bar>caker: did finnix 86.2 get deployed to the other hosts?
20:39<internat>hmm sok, im going to try and modify that a bit, so that it returns 1 or 0 so that i can make sure it doesnt msg me constantly ever 5 minutes until i fix that problem :)
20:41[~]mikegrb would just run it every few hours or some such
20:41<internat>yeah.. well i have a script that checks hosts every 5 minutes and checks that they are still up :)
20:45|-|cow [] has joined #linode
20:45<internat>msgs me when they go down, and msgs me when/if they come back up
21:01|-|womble [~mpalmer@] has joined #linode
21:01|-|Newsome [] has joined #linode
21:12|-|jtickle [] has joined #linode
21:14|-|emcnabb [] has joined #linode
21:28|-|womble [~mpalmer@] has quit [Quit: What's behind the round window...]
21:39|-|spr [] has joined #linode
22:13|-|FireSlash [] has quit [Quit: Leaving]
22:41|-|dc0e [] has quit [Remote host closed the connection]
22:46|-|dc0e [] has joined #linode
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
23:20|-|jtickle [] has quit [Quit: Leaving]
---Logclosed Tue Jan 24 00:00:56 2006