00:21.29 | *** join/#gllug Leeds (~richardc@pcd424154.netvigator.com) |
05:34.31 | *** join/#gllug Leeds (~richardc@www.scorefive.com) |
07:17.10 | morsing | Leeds! |
11:56.18 | *** join/#gllug ibot (~ibot@rikers.org) |
11:56.18 | *** topic/#gllug is Welcome to the Greater London Linux User Group IRC channel | http://www.gllug.org.uk |
11:56.26 | hali | i said that resizing local storage on a box is a bad idea |
11:57.46 | hali | and lol |
11:57.58 | hali | you are talking from an AIX background, which i respect |
11:58.06 | hali | but the topic here is LVM on Linux |
11:59.54 | ChoHag | Personally I would use lvm on a server but would be unlikely to take advantage of it. It's most likely to be there just in case. |
12:01.39 | hali | possibly |
12:01.39 | hali | yes |
12:02.39 | ChoHag | Or on my home servers, of course, which are not run in the same way I'd run 'real' servers. |
12:03.16 | hali | but i would never use Linux LVM on SAN storage assigned to a server, as that would probably be a clustered setup anyway and LVMs cluster support is pretty horrible |
12:03.31 | wethrin | Lustre! |
12:03.32 | ChoHag | No that would be nuts. |
12:03.42 | hali | so the resize question is out of scope as i'd never have anything to expand on |
12:03.47 | ChoHag | LVM for local hardware only. |
12:03.58 | hali | lustre is the nuts, especially if you have developers clever enough to use the lustre api instead of file system calls |
12:03.59 | wethrin | ZFS FTW |
12:04.00 | antiphase | LVM isn't the correct answer when the question is enterprise-grade storage, unless your business isn't in the business of making money |
12:04.14 | ChoHag | Unless you want to do something sick like add a SAN PV to a local VG to bide time until another disc arrives. |
12:04.23 | ChoHag | Though in that case you're well into bad planning territory. |
12:04.24 | hali | zfs lacks cluster support, lustre+zfs in 3 years could be cool |
12:04.45 | antiphase | Cluster file systems have cluster in the name, which is coincident because "clusterfuck" also does, and it's the same thing |
12:05.35 | wethrin | Lustre's fine until it can't talk to your NIS server |
12:05.39 | wethrin | and then it rather shits itself |
12:05.46 | ChoHag | I thought lustre was still in the new-but-pretend-we're-not stage that all open sores projects go through. |
12:06.05 | antiphase | wethrin: Then you wake up from your coma and it's not 1973 any more |
12:06.07 | ChoHag | wethrin: Lustre needs NIS? |
12:06.12 | ChoHag | Will that still not die? |
12:06.36 | wethrin | antiphase: Please, be fair. 1983. |
12:06.49 | wethrin | ChoHag: No, but it needs a user database |
12:08.01 | ChoHag | Oh is there any way to convert a btrfs volume to zfs? |
12:08.05 | ChoHag | btrfs sucks. |
12:08.58 | antiphase | rsync |
12:09.05 | ChoHag | And what would be better at hosting zfs? opensolaris or freebsd? |
12:09.30 | antiphase | Obviously opensolaris, which has been canned, so you shouldn't use it |
12:09.50 | wethrin | FreeBSD ZFS should be stable if you're running on 64-bit hardware |
12:10.04 | wethrin | depends on the size of your disks, really |
12:10.20 | wethrin | it was perfectly stable on one of my machines with a few 73GB hard disks |
12:10.25 | hali | i've given up on cool file systems to be honest |
12:10.48 | wethrin | it was crashing all over the place with another that had 6x 1TB drives in it |
12:10.49 | hali | netapp ftw, if you need block storage you are probably on a db box where you can use either the database's storage manager or just raw volume |
12:10.51 | wethrin | (both 32-bit systems) |
12:11.16 | antiphase | Hitachi HDS also ftw if you're drowning in cash |
12:12.14 | ChoHag | Bigger than that. |
12:12.37 | antiphase | The BBC do use ZFS for storage of iPlayer and associated video in various formats iirc, and they've only totally almost lost all their data a few times |
12:13.00 | ChoHag | 2x2TB and either 2x500GB or 1x500+2x250, depending on whether I raid them together or let zfs deal with that. |
12:13.00 | wethrin | yeah, but that's why you have backups |
12:13.30 | ChoHag | Oh it's nothing important. |
12:13.34 | ChoHag | Mostly $HOME and /Stuff |
12:13.37 | antiphase | I bet restoring petabytes takes a while |
12:13.41 | ChoHag | Where /Stuff contains ... stuff. |
12:13.58 | ChoHag | Also, on a related note, I really fucking hate NFS. |
12:15.25 | hali | i love NFS |
12:15.33 | hali | if a proper NFS implementation is used |
12:15.39 | hali | Linux NFS server is not one |
12:15.45 | ChoHag | Ah, well, no. I use Linux. |
12:15.56 | ChoHag | Although if I use ZFS I probably wouldn't be. |
12:16.04 | ChoHag | Although it'd still be on the client. |
12:16.32 | hali | linux NFS client is ok'ish |
12:16.45 | hali | we use NFS extensivly for pretty much everything |
12:17.03 | hali | oracle db storage, vm storage, shared application caches etc |
12:20.12 | wethrin | All networked file systems syck |
12:22.29 | hali | FCoE! :) |
12:22.53 | hali | unified fabric malarky is thefuture </cisco-marketing> |
12:23.05 | hali | you needz nexus |
12:29.39 | ChoHag | Hahaha. Classic. |
12:30.01 | ChoHag | <Most senior admin> I like asking for the reasons it prevents me from doing stupid and non-important things from time to time |
12:30.07 | ChoHag | I don't think it does... |
12:37.29 | ChoHag | Oh latest entertaining Linux problem at home - kernel 2.6.34 (whose initrd hasn't been touched) no longer detects my sata discs. 2.6.32 on the same box does. |
12:37.57 | ChoHag | This happened after I plugged an apparently broken pci-e graphics card in. |
12:38.31 | ChoHag | Boot. Happy. Reboot and remove gfx card. Sad. |
13:23.26 | ChoHag | Oh-shit-everybody-shops day in the US is on Friday. |
13:23.37 | ChoHag | Today we are discussing sharing the load on the US web servers. |
13:24.01 | *** join/#gllug dick_turpin (~dick_turp@host217-34-163-30.in-addr.btopenworld.com) |
13:25.12 | *** join/#gllug wethrin (dankolb@hades.eco.li) |
13:28.00 | dick_turpin | Afternoon all |
14:21.39 | z00dax | some of the people I am talking about are actually on the gllug list. |
14:21.43 | z00dax | maybe I should stop |
14:21.50 | z00dax | dick_turpin: fix it! |
14:21.59 | z00dax | s/fix/check/ |
14:22.40 | dick_turpin | What's that then? |
14:23.38 | z00dax | beer! |
14:25.46 | dick_turpin | z00dax: I see my mate Dan Lynch (methoddan) is now a fully paid up CentOS fanboy |
14:27.09 | z00dax | dick_turpin: :) |
14:27.15 | z00dax | we need to get him a CentOS coffee mug |
14:28.10 | dick_turpin | He'd love that, I've met him in person loads of times by the way |
14:30.08 | z00dax | :) |
14:31.22 | dick_turpin | He's a big bloke |
14:36.50 | dick_turpin | z00dax: What was you talking about earlier? Is it something on the mailing list? |
14:38.08 | z00dax | dick_turpin: yeah, talking about how its hard to find people with $clue these days on linux/admin roles |
14:38.36 | z00dax | someone was looking to hire 2 guys, and asked me to sit through some of the interviews with them. |
14:38.50 | dick_turpin | Oh yeah just reading it now, so did you really give a guy a chance to jump on a box and google the answer? |
14:41.18 | dick_turpin | Ooh I'd have said 254 that's the ip of my sales box |
14:42.47 | z00dax | yeah, i think the problem is that this specific job specified it needs to be someone with >= 5 years of working with a web based environment |
14:43.03 | z00dax | and many people think sysadmin = sysadmin |
14:44.30 | dick_turpin | Hm remote support? |
14:45.22 | z00dax | more like : worked as the main systems guy for a website that had more than 20 servers |
14:46.14 | dick_turpin | Oh I see, sorry so this is a webhosting company |
14:46.31 | z00dax | no, just a company with a single website that is spread over a bunch of machines |
14:46.50 | dick_turpin | ? An Intranet? |
14:48.40 | *** join/#gllug gethoper (~gethoper@175.217.169.130) |
15:11.31 | z00dax | dick_turpin: some of the busy / larger sites online are spread over quite a few machines. eg. facebook is rumoured to be 50,000 machines or more |
15:11.45 | z00dax | the guys at zynga claim adding a few hundred machines every week |
15:12.25 | dick_turpin | Yeah, I realised the stupid response I gave after I hit send, seems like I won't be getting the £50K job :-( |
15:16.49 | *** join/#gllug wethrin (dankolb@hades.eco.li) |
15:38.08 | z00dax | recommand a UK based DID->voip provider ? |
15:44.51 | dick_turpin | Ooh hang on I think I can |
15:46.52 | z00dax | getting in touch with genzee |
15:47.04 | z00dax | i know gradwell comes up often |
15:47.09 | z00dax | anyone else in the mix ? |
15:47.12 | z00dax | not sipgate |
15:47.49 | dick_turpin | When the stupid mail comes through from my windows box |
15:48.27 | dick_turpin | Paul Harrison |
15:48.29 | dick_turpin | Sales Director |
15:48.31 | dick_turpin | The Voice Factory |
15:48.33 | dick_turpin | 020 3326 0800 |
15:48.35 | dick_turpin | Network House |
15:48.37 | dick_turpin | Basing View |
15:48.39 | dick_turpin | Basingstoke, Hants RG21 4HG |
15:48.41 | dick_turpin | Any use? |
15:50.59 | z00dax | sure, I'll get in touch. thanks |
15:56.02 | *** join/#gllug wethrin (dankolb@hades.eco.li) |
16:00.17 | morsing | wethrin: 4.8? |
16:03.23 | wethrin | yep |
16:08.08 | dick_turpin | wethrin: Doctor! |
16:09.22 | wethrin | Peon! |
16:09.24 | dick_turpin | z00dax: Ooh goody so seeing as I'm the font of knowledge can I have the job please, I can give wethrin as a reference. |
16:21.44 | dick_turpin | thinks this waiting for an interview lark is not much fun |
16:21.55 | dick_turpin | bilarh: ! |
16:22.46 | z00dax | dick_turpin: sure :) |
16:26.04 | dick_turpin | I see the thread just goes on and on :-( |
16:28.42 | z00dax | dick_turpin: the important thing is - are you enjoying it ? |
16:29.34 | dick_turpin | Yes its just a pity I don't understand it ;-) |
17:08.42 | *** part/#gllug dick_turpin (~dick_turp@host217-34-163-30.in-addr.btopenworld.com) |
18:52.30 | *** join/#gllug AcerRevo (~peter@static-87-243-206-72.adsl.hotchilli.net) |
19:28.09 | *** join/#gllug Alex (alexjs@goatse.co.uk) |
21:05.55 | *** join/#gllug Armand (Armand@cpc1-haye3-0-0-cust311.haye.cable.virginmedia.com) |