[02:23] is anyone still running Lucid? or has everyone updated to Maverick? [02:26] Nope, still Lucid on everything here. [02:27] I've got a 64-bit install. [02:27] AMD Athlon X2 5400+ w/8GB RAM, lots of disk [02:27] ran for two years on 8.04.x stably [02:27] I upgraded to 10.04 so I could try out KVM [02:27] and now the thing is hard hanging a couple of times per day [02:27] Weird. [02:27] I've got it loaded with noacpi as a kernel argument now [02:28] but that doesn't seem to help [02:28] nothing gets logged in dmesg or /var/log/messages [02:28] and there is no apparent slowdown before it halts [02:28] Try running a command-line-only system for a few hours and see if anything happens. (To distinguish between a graphics driver problem and something else) [02:28] it _is_ command line only [02:29] headless setup, usually [02:29] oh [02:29] I've got a cheapo 15" lcd plugged into it now, to read the text consoles. [02:29] Well now that ain't good. [02:29] when it dies, it deactivates the VGA too [02:29] and takes my cheapo usb-based KVM switch with it, too [02:29] Anything in /var/crash/ [02:29] ? [02:29] ooh, I'll check [02:29] nope, dead empty [02:29] crud [02:29] /var/log/kern* hasn't a thing either [02:30] /var/log/debug? [02:30] /var/log/faillog? [02:31] debug drops off after xinetd started after the last crash and starts up again with a kernel entry... [02:31] Dec 7 16:38:11 localhost xinetd[2567]: removing time [02:31] Dec 7 19:48:13 localhost kernel: [ 0.000000] e820 update range: 0000000000000000 - 0000000000010000 (usable) ==> (reserved) [02:31] Dec 7 19:48:13 localhost kernel: [ 0.000000] MTRR default type: uncachable [02:31] nope, only two failed logins listed in /var/log/faillog -- root on 06/20/10, and me on 06/25/09 [02:32] this thing is knocking the tar out of me [02:32] Any chance it's actually a hardware problem that coincidentally is happening after the OS upgrade? [02:32] conceivable. [02:32] cpuburn didn't turn up any heating problems [02:32] dang this is my external gateway. [02:33] when I go to bed tonight, I'll bring it down and fire up memtest86+ [02:34] I'm worried that it could be a SATA controller going, or mobo failing. [02:34] after burning 3 days on this, I'm thinking I'm getting to the point where I could throw a new machine at the problem. [02:35] Does your hard drive support SMART? [02:35] yeah, they're all working AOK [02:35] though I don't trust SMART too far. [02:35] 'k [02:36] it's kind of dumb when it comes to predicting errors [02:36] all six satas say they're aok. [02:37] but...it does make me wonder about power... [02:37] I wonder if my PS is acting up. [02:37] You could always read through all of the posts on this thread: http://ubuntuforums.org/showthread.php?t=1478787 [02:38] ;) [02:38] (Although waiting for the next LTS might be faster.) [02:39] I went through the first 25 pages of that, then skipped ahead to the end [02:39] where people say "huh. no solution found." [02:39] mainly because I think there are a lot of different root causes [02:41] https://help.ubuntu.com/community/FaultyHardware might be useful, although it sounds like you already knew these. [02:44] 500watt powersupply for six sata drives...I wonder if I'm pushing my luck here [02:47] Care to list off what else you have in there? (PCI cards, fans, etc.) [02:47] Intel PCI-express NIC [02:47] that's it [02:47] Not a single fan eh? That could be your problem too :P [02:47] a dvd burner that's not even plugged in, because I'm out of sata cables [02:48] hahaha [02:48] =) [02:48] got me. [02:48] okay, fan on the back [02:48] nice big quiet fan on the CPU [02:48] I'd have to look to see if there are others [02:49] I'm seeing posts on forums, telling me that other people consider 6 drives on a single 500w P/S to be pushing the envelope [02:52] jeremyA: what socket is that Athlon? [02:55] AM2+ I think [02:55] am2 [02:55] sitting in a biostar a740G M2+ motherboard [02:55] with 8gb of ddr2 ram [02:56] 65watt Athlon 64X2 5400 brisbane dual-core [02:57] well this calculator doohickey just recommended 312W, so that doesn't seem bad yet. [02:57] wait'll we add in the drives... [02:57] that's with the drives, assuming 7200rpm [02:57] oh, okay. [02:57] hrm. [02:57] should be fine then [02:57] unless the P/S is failing [02:57] but it's usually not dying at periods of particularly heavy load [02:58] and I'm almost never over 4gb RAM usage [02:58] Mem: 8128336 4232644 3895692 0 41628 2135184 [03:00] two of these drives are ancient 160gb 7200rpm barracudas...but I don't think they'd be taking 100w of power each or anything [03:01] where'd you find the calculator, anyway? :) [03:02] google "PSU calculator", first result [03:03] thx [03:07] newegg has one, too: http://educations.newegg.com/tool/psucalc/index.html [03:07] they're suggesting a 408w P/S for me. [03:22] tonyyarusso: are all your machines Intel hardware? [03:22] jeremyA: Nope. I'm typing on an AMD Phenom II X4 905e machine currently. [03:23] okay, good to know [03:23] I've been reading through that thread, and noticed a guy saying "lucid only freezes on my AMD hardware" [03:23] The rest are, unless you count the ARM gadget. [03:25] interestingly, I noted that the BIOS had the ACPI set to to v1. I set it to v2 when it was down last, and deactivated the onboard sound [03:27] I have a board here that's on v2 now, but has an option for v3. [03:27] yeah, I have to take the gateway down to refsck a disk once my wife is done checking email [03:27] but since I'm booted with noacpi, I don't think the acpi version that BIOS is using should matter [03:28] bbiab [03:54] so...how the heck do I enabled the --mark-- in rsyslog? [03:55] I put these lines in /etc/rsyslog.conf [03:55] $ModLoad immark # provides --MARK-- message capability [03:55] $MarkMessagePeriod 60 [03:55] but nowhere in any logfile do I find my MARKs [03:55] whoops, my bad [03:55] there they are in kern.log [03:55] not in messages where I expected them [03:55] * jeremyA <-- dummie [18:57] hey, tonyyarusso, you around? [19:18] jeremyA: kinda - at work, but yes. [19:21] do any of your amd machines use tsc as a clocksource? [19:21] after upgrading to the 2.6.34 kernel, I got a note about my tsc clocksource being unstable. [19:21] so I've set it to use clocksource=hpet on the next book [19:24] um, I have no idea. I've never touched anything like that. [19:24] * tonyyarusso uses default kernels and such