--- Day changed --- Log opened Sun Apr 09 23:59:05 2006 00:03 -!- TheFirst [gaveup@CPE-70-92-72-102.new.res.rr.com] has joined #linode-xenbeta 00:04 < TheFirst> hm server's still down? 00:05 < taupehat> no 00:05 < taupehat> is up here 00:05 < taupehat> brb 00:06 < TheFirst> hmmm 00:07 < TheFirst> node didnt restart after the reboot :/ 00:07 -!- ^GaveUp^ [gaveup@your.friendly.neighborhood.hellmouth.info] has joined #linode-xenbeta 00:07 -!- TheFirst [gaveup@CPE-70-92-72-102.new.res.rr.com] has quit [Quit: ] 00:10 -!- ^GaveUp^ is now known as TheFirst 00:12 < TheFirst> hmmm well better io now....dunno if i'd say it's great, though... Timing buffered disk reads: 16 MB in 3.01 seconds = 5.32 MB/sec 00:19 -!- snorp is now known as snorp|out 00:26 < taupehat> heh 00:26 < taupehat> feels good to me 00:29 < warewolf> that's what SHE said. 00:29 < warewolf> yes, I did just go there 00:29 < taupehat> =] 00:29 < warewolf> and no you can't stop me. 00:29 < taupehat> true 00:29 < taupehat> but caker can! 00:30 < warewolf> true 00:32 < caker> weird .. only a fraction of the nodes started 00:32 < taupehat> heh 00:32 < taupehat> I issued a boot 00:33 < taupehat> no idea whether it would have started without that 00:33 < taupehat> on an odder note 00:33 < taupehat> my cheesy actiontec dsl modem is locking dns stuff to bad addresses 00:33 < taupehat> constantly 00:33 < taupehat> it'll resolve the mailserver I use for work to the address for gmail.com 00:34 < taupehat> and told me that I had taupehat.com at google's IP address 00:34 < taupehat> other non-google-related nuttiness 00:34 < warewolf> er what? 00:34 < warewolf> how is your dsl modem doing dns for you? 00:34 < warewolf> bad 00:34 < warewolf> bad taupehat 00:35 < taupehat> it's not doing dns for long 00:35 < taupehat> this is on my home lan 00:35 < taupehat> I've tried telling the damn thing to use known-good DNS servers, too 00:35 < taupehat> it just sucks 00:37 < TheFirst> speaking of broken dns...my lan dns is broke too 00:38 < warewolf> ... which is why I run my own caching nameserver :) 00:38 < warewolf> so I can hijack a couple of domains, like doubleclick.net and burstnet.net :) 00:40 < taupehat> hehe 00:40 < taupehat> yeah 00:40 < taupehat> I do similar trickery at work 00:41 < taupehat> the dns server insists that *doubleclick.net can be found at 127.0.0.1 00:42 < TheFirst> warewolf: that's how i _should_ be setup 00:43 < taupehat> hehe 00:43 < taupehat> finding problems for this modem with regard to DNS is easy 00:43 < caker> hmm .. I could issue sysrqs to get process lists 00:43 < caker> but not swaps 00:44 < TheFirst> wish i knew dns well enough to make troubleshooting easy 00:44 < caker> TheFirst: what's up? 00:45 < warewolf> # -SSH- warewolf@xabean:~$ cat hitbox.com 00:45 < warewolf> $TTL 86400 00:45 < warewolf> $ORIGIN hitbox.com. 00:45 < warewolf> @ 1D IN SOA @ root ( 42 3H 15M 1W 1D) 00:45 < warewolf> 1D IN NS @ 00:45 < TheFirst> got tinydns/dnscache setup on my lan but dnscache doesn't properly forward lan lookups to tinydns 00:45 < warewolf> # -SSH- warewolf@xabean:~$ 00:45 < warewolf> that's hitbox.com 8-) 00:45 < taupehat> heh 00:45 < TheFirst> although from what i read everything is configured the way it should be 00:46 < taupehat> last time I setup bind here, I ended up publishing www.taupehat.com as 192.168.1.10 00:46 < warewolf> heh 00:46 < taupehat> that was leet sauce 00:46 < warewolf> good lord 00:46 < warewolf> I have 2300+ messages in my Lists/Bugtraq folder 00:47 < TheFirst> can do a say .... dig -x 192.168.0.2 @localhost and that reverses properly, but a dig -x 192.168.0.2 @192.168.0.1 which should forward to 127.0.0.1 for lookup doesn't and consequently fails 00:47 < taupehat> warewolf: that's, what, 3 months worth? 00:51 < warewolf> taupehat: um 00:52 < warewolf> I last 'Mark -> All as read' sometime in november. 00:52 < warewolf> read some in december 00:52 < warewolf> read some in january 00:53 < warewolf> little bit in feb 00:53 < warewolf> and then nothing 01:00 < TheFirst> aha figured out my stupid dns issue 01:01 < TheFirst> now just the even more annoying 2MB/s speed issure 01:20 < linbot> New news from forums: Filesystem options? in Feature Request/Bug Report 01:29 < caker> ahah .. tiobench 01:29 < TheFirst> ? 01:29 < caker> was trying to find that the other day 01:31 < TheFirst> oh wow...back to good hdparm results now... 01:31 < TheFirst> Timing buffered disk reads: 120 MB in 3.03 seconds = 39.62 MB/sec 02:03 < TheFirst> figures...now it's back down to 172K/s 03:55 -!- TheFirst [gaveup@your.friendly.neighborhood.hellmouth.info] has quit [arion.oftc.net kinetic.oftc.net] 03:55 -!- Marcel [~marcel@jharkema.xs4all.nl] has quit [arion.oftc.net kinetic.oftc.net] 03:55 -!- gsf_ [ydTotDsyfF@panix3.panix.com] has quit [arion.oftc.net kinetic.oftc.net] 03:55 -!- tierra [~tierra@ibaku.net] has quit [arion.oftc.net kinetic.oftc.net] 03:55 -!- warewolf [warewolf@warewolf.org] has quit [arion.oftc.net kinetic.oftc.net] 03:55 -!- encode [~encode@blah.i.hate.w1ndo.ws] has quit [arion.oftc.net kinetic.oftc.net] 03:55 -!- gpd [~gpd@70.85.16.173] has quit [arion.oftc.net kinetic.oftc.net] 03:55 -!- caker [~caker@caker.netrep.oftc.net] has quit [arion.oftc.net kinetic.oftc.net] 03:55 -!- linbot [~supybot@ns.theshore.net] has quit [arion.oftc.net kinetic.oftc.net] 03:56 -!- TheFirst [gaveup@your.friendly.neighborhood.hellmouth.info] has joined #linode-xenbeta 03:56 -!- Marcel [~marcel@jharkema.xs4all.nl] has joined #linode-xenbeta 03:56 -!- gsf_ [ydTotDsyfF@panix3.panix.com] has joined #linode-xenbeta 03:56 -!- gpd [~gpd@70.85.16.173] has joined #linode-xenbeta 03:56 -!- encode [~encode@blah.i.hate.w1ndo.ws] has joined #linode-xenbeta 03:56 -!- warewolf [warewolf@warewolf.org] has joined #linode-xenbeta 03:56 -!- linbot [~supybot@ns.theshore.net] has joined #linode-xenbeta 03:56 -!- caker [~caker@caker.netrep.oftc.net] has joined #linode-xenbeta 03:56 -!- tierra [~tierra@ibaku.net] has joined #linode-xenbeta 03:56 -!- mode/#linode-xenbeta [+o caker ] by kinetic.oftc.net 09:12 -!- bendy24 [~scott@dsl-216-110-238-70.wightman.ca] has joined #linode-xenbeta 09:38 -!- snorp|out is now known as snorp 09:54 -!- sednet [~86920016@webuser.linode.com] has joined #linode-xenbeta 10:00 < sednet> Anyone else getting really bad performance from host56? 10:02 < sednet> Microsoft general solution number 1 is it. 10:02 < mikegrb> ha ha 10:02 < snorp> sednet: yes. 10:03 < snorp> sednet: have been for quite a while :( 10:03 < mikegrb> I never would have believed how many people just reboot their linode when preformance was bad if I hadn't seen it for myself 10:03 < snorp> but I think caker, et al, is a ware 10:03 < snorp> aware, even 10:03 < sednet> Strangely it seems to help 10:03 < mikegrb> especially humerous because it just makes it worse 10:03 < snorp> it actually does seem to help for some reason 10:03 < snorp> like it's all responsive and stuff after a reboot 10:03 < mikegrb> it doesn't 10:04 < snorp> it totally does 10:04 < sednet> It helps for 12 hours or so. No doubt it's not good for the host 10:04 < snorp> I can ssh in, for instance :) 10:04 < snorp> right 10:04 < mikegrb> right and I've got a billion in cash in my pocket 10:05 * bendy24 picks mikegrb's pocket 10:05 < bendy24> $.50! 10:05 < bendy24> cant even go to mcdonalds 10:05 < bendy24> pfft 10:05 < mikegrb> anyway, when I see people do that, I throttle their io conciderably and renice their Linode so they can't hurt the other people on their host 10:06 < snorp> eek 10:06 < mikegrb> and it's normally brought to my attention by other people complaining that their host is slow 10:06 < taupehat> hehe 10:06 < snorp> maybe that's why host56 is so slow 10:06 < snorp> people constantly rebooting 10:06 < taupehat> I haven't rebooted (except when the host does) and my performance is also suboptimal 10:07 < snorp> yeah, I haven't rebooted other than then either 10:07 < snorp> and I can't even ssh in 10:08 < bendy24> i had to reboot once to get networking working 10:08 < bendy24> since the last reboot that is 10:08 < taupehat> so how does io throttling work on the xenodes? 10:11 < taupehat> hehe 10:12 < taupehat> whatever you're doing to throttle IO, it's working! 10:12 < taupehat> taupehat:~# hdparm -t /dev/hda1 10:12 < taupehat> /dev/hda1: Timing buffered disk reads: 2 MB in 56.55 seconds = 36.21 kB/sec 10:12 < taupehat> I think that's about how fast a floppy reads 10:13 < snorp> that is way slower than a floppy I think 10:14 < taupehat> and this is me not rebooting my linode, mikegrb 10:14 < snorp> I think that's about half the speed of a floppy :) 10:14 < taupehat> did you just test it? 10:15 < snorp> no. 10:15 < snorp> oh, I could be wrong 10:16 < taupehat> hrm 10:16 < snorp> it might be 1/4 the speed of a floppy 10:16 < snorp> anyway, it's freakin slow :) 10:16 < taupehat> it takes about 45 seconds to read all the data off a 1.44M floppy if I remember right 10:16 < taupehat> so 2M in 56 seconds fits 10:16 < taupehat> or close enough for the gripe =] 10:16 < snorp> heh 10:17 < taupehat> s/gripe/whine 11:01 < sednet> Performance still sucks but it's better than before 11:02 < sednet> mikegrb - did you knacker my xenode or are they all this slow? 11:20 < linbot> New news from forums: Bridging between wired and wireless interfaces in Linux Networking 11:32 < bendy24> anyone notice better performance? 11:35 -!- snorp [~snorp@CPE-65-28-73-215.kc.res.rr.com] has quit [Read error: Connection reset by peer] 11:35 < sednet> No. Still won't run a PHP webpage 11:37 < bendy24> gah 11:37 < sednet> It's impressively bad. The disks can't be nearly up to the task 11:38 < bendy24> and its mirrored.. 11:38 < bendy24> i wonder how many domains there are now 11:38 < sednet> RAID-1 makes writtes slower, not faster 11:38 < sednet> Only reads are faster 11:38 < bendy24> true 11:40 < sednet> We need a raid 5 of 15k disks with a really large cache before them 11:40 < sednet> Or less domains on the host 11:41 < bendy24> does raid 5 help much? 11:41 < TheFirst> and yet that still wouldnt solve these issues 11:41 < sednet> It would spread the disk load if thats the problem 11:41 < sednet> If this is a software not a hardware problem it won't help much 11:42 < TheFirst> bingo 11:44 < mikegrb> writes are faster in this case 11:44 < mikegrb> with write cachehing 11:48 < taupehat> owch 11:52 -!- ExquisitePope is now known as NonVolatileRambo 11:58 < taupehat> /dev/md0: Timing buffered disk reads: 214 MB in 3.01 seconds = 71.14 MB/sec 11:59 < taupehat> that's my raid array 12:00 < taupehat> not bad for sata disks running raid5 under software 12:00 < NonVolatileRambo> your Redundant Array of Independent Disks Array 12:01 < taupehat> ... 12:01 < NonVolatileRambo> it's a Matrix of disks! 12:01 * taupehat beats NonVolatileRambo over the head with a NIC card 12:01 * NonVolatileRambo smashes taupehat into the ATM machine 12:01 * taupehat cats NonVolatileRambo into his PCI interface 12:01 < sednet> Cut it out you two 12:02 * NonVolatileRambo cuts out taupehat's PIN number 12:02 < taupehat> ... 12:02 < taupehat> actually 12:02 < NonVolatileRambo> it's tattooed on his nipple 12:02 < taupehat> it's a JBOD array running RAID 5 dropped into LVM 12:02 < taupehat> chew on that for a while =] 12:02 < NonVolatileRambo> I had one of those once 12:02 < NonVolatileRambo> too noisy 12:02 < NonVolatileRambo> I shut it down and moved to linode 12:02 < taupehat> I still think it's decent performance for what it is 12:03 < sednet> JBOD running RAID 5? What on earth are you on about? 12:03 < taupehat> heh 12:03 < taupehat> ... 12:03 < NonVolatileRambo> haha 12:03 < NonVolatileRambo> the JBOD isn't doing the "running" 12:03 < NonVolatileRambo> by definition 12:03 * taupehat thinks it's too early in the week to be kicking people in the shins, but he may be forced to make an exception 12:03 < taupehat> or two... 12:03 < sednet> It's not a JBOD then. It's software RAID 5 12:04 < taupehat> sigh 12:04 < taupehat> get technical, willya 12:04 < NonVolatileRambo> the hardware is still called a JBOD 12:04 < NonVolatileRambo> even if the software gangs 'em together 12:05 < sednet> No it's not. 12:05 < NonVolatileRambo> Yes, it is. 12:05 < sednet> Dumbarse helpdesk scum 12:06 * NonVolatileRambo sets sednet on fire 12:07 < sednet> It's a JBOD if the disks are appended. 12:08 < taupehat> sednet: how's this 12:09 < taupehat> I'm getting 3-packs of hard drives, dropping them into this 15-disk chassis, and making raid5s out of those 3-packs, then adding them to the LVM 12:09 < taupehat> so technically, it's a JBOD of raid5s 12:10 < NonVolatileRambo> aren't they giving that goofy names like "RAID 51" or "RAID 15" or something? 12:10 < taupehat> nfi 12:10 < taupehat> I just do what works for the task 12:10 < taupehat> kickass chassis, by the way 12:11 < taupehat> this one: http://supermicro.com/products/chassis/3U/933/SC933T-R760.cfm 12:36 -!- sednet [~86920016@webuser.linode.com] has quit [Quit: CGI:IRC] 13:54 -!- pclissold [~pclissold@sassenheim.clissold.nl] has joined #linode-xenbeta 14:26 < bendy24> go xen go 14:26 < TheFirst> ooo new low... Timing buffered disk reads: 2 MB in 85.13 seconds = 24.06 kB/sec 14:27 < bendy24> gimme a x 15:38 -!- Marcel [~marcel@jharkema.xs4all.nl] has quit [Quit: Leaving] 16:33 -!- marc_in_lux [~marc@cable-83.217.130.243.coditel.net] has joined #linode-xenbeta 16:47 -!- bendy24 [~scott@bendy.netop.oftc.net] has quit [Quit: leaving] 16:58 -!- pclissold [~pclissold@sassenheim.clissold.nl] has left #linode-xenbeta [] 17:13 < linbot> New news from forums: No reply to upgrade ticket 19075 in Sales Questions and Answers 17:14 -!- marc_in_lux [~marc@cable-83.217.130.243.coditel.net] has quit [Quit: Leaving] 18:16 < TheFirst> Timing buffered disk reads: 2 MB in 141.83 seconds = 14.44 kB/sec 18:16 < TheFirst> yay a new low 19:20 < caker> geesh 19:41 < TheFirst> ? 19:44 < TheFirst> ooo yay i've improved! 19:44 < TheFirst> Timing buffered disk reads: 2 MB in 132.21 seconds = 15.49 kB/sec 19:44 < NonVolatileRambo> haha 19:44 < NonVolatileRambo> are you running bonnie from your node? 19:46 < TheFirst> nope, completly idle 19:46 < NonVolatileRambo> what's the tool you're getting that from? 19:47 < TheFirst> hdparm 19:47 < NonVolatileRambo> ah, right 19:54 < fo0bar> -Tt 20:35 -!- sprouse [sprouse@chewbacca.infonurse.net] has quit [Quit: installed 5.10] 20:46 -!- egatenby [~egatenby@itsb149.itsnpt.com] has joined #linode-xenbeta 20:54 -!- sunny [user@lexington.opencurve.org] has quit [Quit: leaving] 21:22 < linbot> New news from forums: swap question in Performance and Tuning 21:35 -!- snorp [~snorp@CPE-65-28-73-215.kc.res.rr.com] has joined #linode-xenbeta 21:43 < snorp> hey, my xenode paniced 21:43 -!- Newsome [~sorenson@adsl-75-0-136-70.dsl.chcgil.sbcglobal.net] has joined #linode-xenbeta 21:43 < snorp> caker, mikegrb: are you guys interested in the trace 21:44 < caker> snorp: I copied it, thanks 21:44 < caker> snorp: do anything to trigger it, or just normal usage? 21:44 < snorp> caker: just normal usage 21:45 < caker> ok 21:45 < snorp> I run apache/mysql 21:45 < snorp> and that's pretty much it 21:45 < snorp> named 21:47 < TheFirst> here's one for you too 21:47 < TheFirst> hellmouth:/home/gaveup# hdparm /dev/hda1 21:47 < TheFirst> /dev/hda1: readonly = 0 (off) readahead = 256 (on) HDIO_GETGEO failed: Inappropriate ioctl for device 21:48 < TheFirst> ok so that didn't copy right 21:48 < caker> yeah -- not surprising 21:48 < TheFirst> lets try this... 21:48 < TheFirst> /dev/hda1: readonly = 0 (off) readahead = 256 (on) HDIO_GETGEO failed: Inappropriate ioctl for device 21:48 < TheFirst> hellmouth:/home/gaveup# 21:48 < TheFirst> hmm ok well there should be newlines in there ;) 21:57 < TheFirst> hmm i suppose a reboot is the best shot at fixing that error? 22:23 -!- egatenby [~egatenby@itsb149.itsnpt.com] has left #linode-xenbeta [] 22:43 -!- egatenby [~egatenby@itsb149.itsnpt.com] has joined #linode-xenbeta 22:46 -!- egatenby [~egatenby@itsb149.itsnpt.com] has left #linode-xenbeta [] --- Log closed Mon Apr 10 23:59:01 2006