| --- Log opened Wed Feb 16 10:45:14 2011 |
| 10:45 | | TheWatcher [chris@Nightstar-b4529b0c.zen.co.uk] has joined #code |
| 10:45 | | Irssi: #code: Total of 26 nicks [7 ops, 0 halfops, 0 voices, 19 normal] |
| 10:45 | | mode/#code [+o TheWatcher] by Reiver |
| 10:46 | | Irssi: Join to #code was synced in 53 secs |
| 12:03 | | Rhamphoryncus [rhamph@C06FE3.F5723C.BE3FEB.9D4666] has quit [Client exited] |
| 12:04 | | celticminstrel [celticminst@Nightstar-f8b608eb.cable.rogers.com] has joined #code |
| 12:07 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has joined #code |
| 12:24 | | AnnoDomini [annodomini@F67919.F326B3.98D923.BDA7B6] has quit [[NS] Quit: Going home.] |
| 12:56 | | AnnoDomini [annodomini@Nightstar-6020b897.adsl.tpnet.pl] has joined #code |
| 12:56 | | mode/#code [+o AnnoDomini] by Reiver |
| 12:57 | | SmithKurosaki [smith@Nightstar-7820a96a.dsl.teksavvy.com] has quit [Ping timeout: 121 seconds] |
| 13:01 | | SmithKurosaki [smith@Nightstar-7820a96a.dsl.teksavvy.com] has joined #code |
| 13:04 | | Tarinaky [Tarinaky@Nightstar-f349ca6d.plus.com] has joined #code |
| 13:29 | | celticminstrel [celticminst@Nightstar-f8b608eb.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 14:52 | | * jerith sighs. |
| 14:53 | <@jerith> | A security release of MySQL seems to make my EC2 instances go on an extended holiday when I do large SELECT queries. :-/ |
| 14:53 | <@jerith> | I have no idea why. I'm not sure how large "large" is. |
| 14:55 | <@jerith> | There are all sorts of variables, but the security point release is the only one that I can't feasibly vary. :-/ |
| 14:55 | < Tamber> | ...oh dear. |
| 14:55 | <@jerith> | (Upgrading an older box risks breaking it and trying to downgrade a newer box results in the package manager whining about not finding stuff.) |
| 14:57 | | * Tamber prods jerith with the dilemma fork. |
| 14:57 | <@jerith> | This is bloody annoying, because I actually need to work on some of this stuff. |
| 14:59 | <@jerith> | I can't even break up the long-running operations, because they're huge and ugly and complicated and were written by someone else. |
| 14:59 | < Tamber> | Ick. |
| 14:59 | <@jerith> | And without being able to run both sets next to each other in a controlled environment, I can't guarantee that I haven't screwed up something important. |
| 15:00 | <@jerith> | Also, breaking them up isn't really all that feasible. |
| 15:01 | <@jerith> | It's stuff like "do this processing on each of the eleventy zillion rows in this table, joined against those other twelve tables using rules cribbed from the Necronomicon". |
| 15:02 | < Namegduf> | Maybe you should just schedule the job to be run at midnight, then. |
| 15:02 | < Tamber> | Oh, so it's /you/ breaking the universe like that! |
| 15:02 | < Namegduf> | I think that's the normal procedure. |
| 15:02 | <@jerith> | Namegduf: It already does. |
| 15:02 | <@jerith> | I need to add features. |
| 15:02 | < Namegduf> | Hm. |
| 15:03 | <@jerith> | Actually, I'm in the process of replacing the entire data loading stage wholesale. |
| 15:03 | <@jerith> | But I need to finish that so I can add features. |
| 15:04 | <@jerith> | I can't do my testing with a smaller dataset than production, because the production dataset has more edge cases than the non-Euclidean parts of R'lyeh. |
| 15:05 | <@jerith> | Back when I tried, I ended up rewriting the damned thing every time I put it through the full set anyway. |
| 15:29 | | AnnoDomini [annodomini@Nightstar-6020b897.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 18:07 | | TarinakyKai [Tarinaky@Nightstar-f349ca6d.plus.com] has joined #code |
| 18:09 | | Tarinaky [Tarinaky@Nightstar-f349ca6d.plus.com] has quit [Operation timed out] |
| 18:15 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has joined #code |
| 18:20 | | Kindamoody|out is now known as Kindamoody |
| 19:18 | | Derakon [chriswei@Nightstar-cfae48c3.ca.comcast.net] has joined #code |
| 19:18 | | mode/#code [+o Derakon] by Reiver |
| 19:18 | <@Derakon> | Eternal fun times at Derakon's workplace! |
| 19:19 | <@Derakon> | One of our Windows 2000 computers gave up the ghost today. |
| 19:19 | <@Derakon> | Bluescreen on boot with "inaccessible boot device" error. |
| 19:19 | <@Derakon> | These computers have no CD drive and cannot boot off of USB. They do have floppy drives, but who the hell uses floppies these days? |
| 19:19 | <@Derakon> | I certainly can't find any W2k installation floppies lying around, which doesn't jibe with our maniacal hording policy. |
| 19:20 | <@Derakon> | (I did find a floppy that held printer drivers, and tried turning it into a recovery floppy with one of our other W2k comps, but it is unable to format the thing, who knows why) |
| 19:21 | <@ToxicFrog> | Boot grub from a floppy, chainload the CD from that? |
| 19:21 | <@ToxicFrog> | Oh wait. No CD drive. |
| 19:21 | <@ToxicFrog> | Sorry. |
| 19:22 | <@Derakon> | Yeah, these are ancient 1U rackables. |
| 19:23 | <@Derakon> | (First response from my boss when I tell him that the thing broke: "It's not like these computers are ten years old or anything." Sadly, he was entirely serious) |
| 19:28 | | Vornucopia [NSwebIRC@C888DE.7F9621.E9EB68.6382E5] has joined #code |
| 19:28 | <@Derakon> | Hey Vorn. |
| 19:28 | | * Vornucopia wavels |
| 19:29 | | * Derakon points Vorn to http://paste.ubuntu.com/567807/ |
| 19:30 | < celticminstrel> | Vornucopia, eh... |
| 19:33 | < Vornucopia> | Der: D: |
| 19:33 | | * Vornucopia is more Vorn than your body has room for. |
| 19:34 | < TarinakyKai> | My BIOS lost all its settings this afternoon when I woke up. |
| 19:34 | < TarinakyKai> | Which was Fun. |
| 19:34 | <@Derakon> | While I'm working on this dead comp, I'm also wondering how in blazes I'm going to run my planned HD backup system if I can't boot these computers into Linux. |
| 19:38 | < Vornucopia> | Pop the drive is your best bet, sadly. |
| 19:38 | <@Derakon> | Ick. |
| 19:38 | < celticminstrel> | Pop? |
| 19:39 | <@Derakon> | That's not conducive to doing this on a regular basis. :\ |
| 19:39 | <@Derakon> | Physically remove. |
| 19:39 | < Vornucopia> | SPeaking of popping a drive I need to find a damn t10 so I can finish getting into the old imac. |
| 19:41 | <@jerith> | I have a 12" iBook that boots as far as a baby blue screen with a mouse cursor on it. |
| 19:41 | <@jerith> | It started doing this a week or two ago. :-/ |
| 19:42 | < Vornucopia> | D: |
| 19:42 | <@jerith> | I got it second hand, so I have no installation media. |
| 19:42 | < Vornucopia> | Yeah, mine won't even go into target drive mode. |
| 19:42 | <@jerith> | It also has no usable battery. |
| 19:42 | <@jerith> | Hrm. Maybe I should try that. But not tonight. |
| 19:47 | <@jerith> | http://gengnosis.blogspot.com/2007/01/level-triggered-and-edge-triggered.html |
| 19:56 | <@ToxicFrog> | Derakon: slap a tiny linux partition on the end of the drive, configure bootloader to boot into windows by default? |
| 19:56 | <@Derakon> | Part of my trouble here is that I'm not remotely a sysadmin. |
| 19:57 | <@Derakon> | But that does sound like it would work. |
| 19:57 | <@Derakon> | Of course, it'd require me to have partitioned the thing ahead of time, so it's no good right now. |
| 19:57 | <@Derakon> | And is it possible to non-destructively partition a drive? |
| 19:58 | <@Derakon> | I don't relish the thought of having to partition and then reinstall everything on the Windows side. |
| 19:59 | < Vornucopia> | Yeah, gparted and kin do that. |
| 19:59 | <@Derakon> | That's assuming I can get into Linux in the first place... |
| 19:59 | <@Derakon> | Our local sysadmin, who is sadly on vacation right now, has a USB DVD drive...no idea if we can boot off of that though. |
| 19:59 | <@Derakon> | (And if we can, then there's no need to partition, since we can just borrow that when we need to backup the drives) |
| 20:00 | <@ToxicFrog> | Nondestructive repartitioning has been the norm for at least five years, yes. |
| 20:00 | <@Derakon> | This is a ten-year-old computer~ |
| 20:00 | <@Derakon> | (But I take your point) |
| 20:01 | <@ToxicFrog> | That said, at this point I do wonder if it would be easier to just have the backup server remote-mount the systems in question and slurp down what it needs. |
| 20:01 | <@Derakon> | We can certainly access the computers over VNC. |
| 20:01 | <@Derakon> | Since that's how we usually work with them. |
| 20:03 | <@Derakon> | Anyway, must get food. BBL. |
| 20:03 | | Derakon is now known as Derakon[AFK2] |
| 20:07 | < Vornucopia> | But with No Boot Device you're not going to be getting in to get the thing going unless you have an ethernet bootstrap floppy. Which would be Scary. |
| 20:23 | | Derakon[AFK2] is now known as Derakon |
| 20:32 | <@Derakon> | Just realized the computers do in fact have CD drives, so... |
| 20:41 | | AnnoDomini [annodomini@Nightstar-ae40b7e9.adsl.tpnet.pl] has joined #code |
| 20:41 | | mode/#code [+o AnnoDomini] by Reiver |
| 20:44 | | * Derakon fires up a Linux rescue CD, starts a smartctl HD diagnostic test. |
| 20:44 | <@Derakon> | Assuming that passes, I guess the next step is to try to get a Win2k recovery floppy or CD, and if that fails, then to clone one of the computer's siblings' drives over. |
| 20:47 | < Vornucopia> | Arg. Why do I struggle so much trying to separate these things. |
| 20:51 | <@jerith> | Because OCD? |
| 20:52 | < Vornucopia> | Probably. |
| 20:52 | <@jerith> | What are you separating? |
| 20:53 | < Vornucopia> | www.funkyhorror.net/vornicus/multiplication_drill.html |
| 20:54 | < Vornucopia> | I'm trying to tease out the multiplication-specific portions so I can farm off the non-specific portions to something that will work for stuff like addition and gcd/lcm and... |
| 20:56 | < Vornucopia> | As you can see from the source it's all kind of glommed together in there right now. |
| 20:57 | <@AnnoDomini> | You got 50 out of 50, for a grade of 100%, in 164.4 seconds. |
| 20:57 | <@AnnoDomini> | :D |
| 20:57 | < Vornucopia> | Heh. When I do it I usually either turn it up to -20 to 20, or cut the time down to 1 minute. |
| 20:58 | < Vornucopia> | (I usually get 49 because i miss stuff like 4*0) |
| 21:10 | <@ToxicFrog> | Derakon: disktest is your tiny filesystem-recovering god. |
| 21:10 | <@ToxicFrog> | Er, testdisk |
| 21:12 | <@Derakon> | Will keep that in mind. |
| 21:12 | <@Derakon> | smartctl is what came up in Google. |
| 21:13 | <@Derakon> | If diagnostics pass then we just have a randomly-corrupted Windows system file, in which case the likely next step would be to try to find a Win2k recovery disk, I guess. |
| 21:14 | | * Derakon tries to remember if drive failing => drive cannot ever be used, or if it's drive failing => drive must be reformatted. |
| 21:14 | < Vornucopia> | Could be either, depending. |
| 21:15 | < Vornucopia> | Or it could be "Drive needs a small fix and then you're back to normal" |
| 21:15 | <@Derakon> | What kind of small fix? |
| 21:16 | < Vornucopia> | Depends on the specific failure. |
| 21:17 | <@Derakon> | All I know at this point is that Windows is unable to boot. |
| 21:17 | <@Derakon> | smartctl's test should be done in ~25 minutes at which point I hope to have more information. |
| 21:17 | <@Derakon> | It could be as simple as Windows' disk driver being corrupt. |
| 21:20 | | Kindamoody [Kindamoody@Nightstar-4764665d.tbcn.telia.com] has quit [Client exited] |
| 21:21 | | Kindamoody [Kindamoody@Nightstar-4764665d.tbcn.telia.com] has joined #code |
| 21:21 | <@Derakon> | (Of course, one then wonders how said driver became corrupt in the first place) |
| 21:23 | < Vornucopia> | Yeah. Corruption, bad blocks, that sort of thing can be worked around, usually. Dead controllers, crashed heads, etc, would be more problematic, but you'd be able to detect those by ear. |
| 21:23 | <@McMartin> | Derakon: If you're running Vista or later, your driver ought to be *signed* and you should be able to check its status under Properties |
| 21:23 | <@Derakon> | Windows 2000, mate. |
| 21:24 | <@McMartin> | Damn. |
| 21:24 | <@Derakon> | Yeah. |
| 21:24 | <@ToxicFrog> | Derakon: yeah, what I mean is, smartctl is "is this drive failing"; testdisk is "shit has gone down and the superblock/partition table/etc is gone, how do I get it back" |
| 21:24 | <@ToxicFrog> | So if you need to get data off this thing, it may be useful to you |
| 21:24 | <@Derakon> | Roger that. |
| 21:25 | | * Derakon eyes TestDisk's description on the intertrons. |
| 21:25 | <@Derakon> | This is not what I would expect a program with "test" in its name to do. |
| 21:25 | <@Derakon> | I would name this "FixDisk". |
| 21:27 | <@ToxicFrog> | Yeah. |
| 21:36 | <@Derakon> | Blugh, I don't like having to maintain computers. |
| 21:37 | <@Derakon> | Especially not decrepit ones. |
| 21:53 | <@Derakon> | Okay, smartctl reports a read failure. |
| 21:53 | <@Derakon> | Which the net suggests means the drive has a bad sector which could be fixable by formatting. |
| 21:54 | <@Derakon> | So I guess use dd to dump the current hard drive contents on the off-chance I need them, then format, then use dd again to clone a sibling computer's drive onto this one? |
| 21:55 | <@Derakon> | Does that sound plausible? |
| 21:55 | < Vornucopia> | I'd try testdisk first, it seems like it should be able to fix stuff like that. |
| 21:56 | <@Derakon> | Fairynuff. |
| 21:56 | < Vornucopia> | time for more work. home in 3 hours or so. |
| 21:57 | | Vornucopia [NSwebIRC@C888DE.7F9621.E9EB68.6382E5] has quit [[NS] Quit: Page closed] |
| 22:08 | | * Derakon finds his notes from the last time he backed up a hard drive over the local network. |
| 22:08 | <@Derakon> | "Used Linux liveCD fr. Eric 2/ bckup Nano." |
| 22:08 | <@Derakon> | Great, Derakon. Good note-taking there. |
| 22:08 | <@Derakon> | No indication of how you did this at all. |
| 22:18 | <@ToxicFrog> | Probably either clonezilla, or the classic dd + ssh combo |
| 22:20 | <@Derakon> | Yeah, but I need to get network access on the liveCD side. |
| 22:20 | <@Derakon> | And remember how to use SSH to connect to the destination computer. |
| 22:22 | <@ToxicFrog> | Network access should happen automatically unless you have an exceptionally weird setup. |
| 22:22 | <@ToxicFrog> | ssh is just ssh user@host |
| 22:22 | <@ToxicFrog> | Something like: dd if=/path/to/drive bs=4M | ssh user@backup-server dd of=/path/to/backup.img bs=4M |
| 22:23 | <@Derakon> | Booting off the liveCD I don't start with network access. |
| 22:23 | <@Derakon> | It says to run netconfig or dhcpcd or pump to get network access. |
| 22:23 | <@Derakon> | And I'm not really clear on how they work. |
| 22:23 | <@Derakon> | I wish I'd taken better notes 10 months ago when I last did this. |
| 22:24 | <@Derakon> | (And by "remember how to use SSH" I meant more "what my account settings are on the target computer"; I know how to use SSH) |
| 22:32 | <@ToxicFrog> | ...what sort of liveCD is this? |
| 22:34 | <@ToxicFrog> | But, in any case, try 'ifconfig' to list network interfaces and then 'dhcpcd interface' to request an address |
| 22:38 | <@Derakon> | I got it working using netconfig and copying the network information from one of our other computers. |
| 22:38 | <@Derakon> | And this is an RIPLinuX CD that I got from a coworker. No idea how old it is. |
| 22:39 | <@Derakon> | I should probably make my own. |
| 22:39 | | * Derakon looks at TF's dd example. |
| 22:39 | <@Derakon> | Why set the blocksize to 4M? |
| 22:40 | <@ToxicFrog> | Because dd's default blocksize is something absurdly tiny like 64k and the overhead of making all of those tiny system calls murders performance dead. |
| 22:40 | <@Derakon> | Ah. |
| 22:41 | <@ToxicFrog> | Also, you should just be able to grab a recent Knoppix liveimage and have something useful. |
| 22:42 | <@Derakon> | ...okay, not sure why I thought dd would be able to read the drive properly when it has a bad sector. Durr. |
| 22:42 | <@ToxicFrog> | conv=noerror |
| 22:42 | <@Derakon> | "I should back up the hard drive before I try repairing it!" |
| 22:42 | <@Derakon> | That will keep errors from stopping it? |
| 22:43 | <@ToxicFrog> | Yes. It will retry a few times, and if that fails it'll just fill that sector in with zeroes in the output and move on. |
| 22:43 | <@Derakon> | Cool, danke. |
| 22:56 | <@ToxicFrog> | Also, if you have 'dcfldd' available, consider using that - it's a more sophisticated version of dd originally designed for forensics work. |
| 23:00 | <@Derakon> | Well, dd's already running, so I won't disturb it. |
| 23:00 | <@Derakon> | Also, it is getting a ton of read/write errors. |
| 23:02 | <@Derakon> | (Thanks for all your help, by the way. I'm completely at sea with this sysadmin stuff) |
| 23:02 | <@Derakon> | (However, I am Taking Notes) |
| 23:16 | <@ToxicFrog> | No problem. |
| 23:21 | <@ToxicFrog> | Also, from what you've said, I would not bet money on being able to resurrect that drive~ |
| 23:25 | <@Derakon> | Even with reformatting? |
| 23:25 | <@Derakon> | (Does using dd to clone a working computer's drive onto this drive count as reformatting?) |
| 23:28 | <@ToxicFrog> | It should - AIUI it what happens is that the drive has a certain amount of "extra" sectors, and when a write attempt fails, it internally remaps the bad sector to one of the extras. |
| 23:28 | <@ToxicFrog> | That said, I have never observed this to actually work as designed. |
| 23:28 | | * Derakon nods. |
| 23:30 | <@ToxicFrog> | (and my guess for that is that if you only have one or two bad sectors, it happens automatically and all you get is a SMART warning if that; by the time you're getting widespread IO errors the drive is hosed) |
| 23:30 | <@ToxicFrog> | No harm in trying, though. |
| 23:31 | <@Derakon> | Right. |
| 23:39 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 23:39 | <@Derakon> | Hrm, dd doesn't specify how many times it tries to read a sector (and thus how many times it'll spew errors out), so it's hard to tell exactly how many sectors are faulty. |
| 23:40 | <@Derakon> | I neglected to capture its output for later perusal, sadly. |
| 23:42 | | shade_of_cpux is now known as cpux |
| 23:43 | | RichardBarrell [mycatverbs@Nightstar-f68eb197.cable.virginmedia.com] has joined #code |
| 23:43 | <@ToxicFrog> | AughWTF |
| 23:44 | <@ToxicFrog> | So I've been using bash for around 20 years |
| 23:44 | <@ToxicFrog> | And in all that time I never noticed the 'help' command |
| 23:44 | <@ToxicFrog> | Fffffffffff |
| 23:44 | | * Derakon snerks. |
| 23:44 | | * Derakon eyes that help. |
| 23:44 | <@Derakon> | This is...not intended for newbies. |
| 23:45 | <@ToxicFrog> | No. It's the help for bash builtin commands. |
| 23:45 | <@ToxicFrog> | It's designed as a quick reference for people who already comfortable with the basics of the Bourne shell. |
| 23:45 | <@Derakon> | Right, but a newbie who happens to be on bash and types "help" (as newbies are wont to do) may be subject to a certain amount of dismay. |
| 23:46 | <@ToxicFrog> | As an alternative to going 'man bash' and scrolling down 2500 lines to SHELL BUILTIN COMMANDS. |
| 23:46 | <@Derakon> | Clearly you should have aliased "man bash " or something along those lines. |
| 23:46 | <@Derakon> | (Does that even work?) |
| 23:48 | | Rhamphoryncus [rhamph@C06FE3.F5723C.BE3FEB.9D4666] has joined #code |
| 23:48 | <@ToxicFrog> | (no) |
| 23:50 | <@ToxicFrog> | (however, env LESS="+/^SHELL BUILTIN COMMANDS" man bash does) |
| 23:51 | <@ToxicFrog> | (more generally, you can define 'man PAGE SECTION' as 'env LESS="+/^$SECTION" man "$PAGE"') |
| 23:53 | < simon_> | your sentences look like lisp. |
| 23:54 | <@ToxicFrog> | It happens. |
| 23:54 | <@Derakon> | IRC convention; lines enclosed in parens are sub-topics and/or asides. |
| 23:54 | <@Derakon> | Or if it's not an IRC convention, it's at least an "us" convention. |
| 23:55 | <@Derakon> | Anyway, time to vanish into the aether. Thanks again for the help, TF and Vorn. |
| 23:55 | | Derakon [chriswei@Nightstar-cfae48c3.ca.comcast.net] has quit [[NS] Quit: leaving] |
| 23:59 | < RichardBarrell> | simon_: don't point it out so often as to make him too self-conscious. ToxicFrog postfix sentences in constructing capable_of is. |
| --- Log closed Thu Feb 17 00:00:02 2011 |