00:08.59 | *** join/#gllug lejocelyn_ (~lejocelyn@host86-173-4-249.range86-173.btcentralplus.com) |
00:41.17 | *** join/#gllug lejocelyn_ (~lejocelyn@host86-173-4-249.range86-173.btcentralplus.com) |
00:45.57 | *** join/#gllug lejocelyn_ (~lejocelyn@host86-173-4-249.range86-173.btcentralplus.com) |
01:02.57 | *** join/#gllug lejocelyn_ (~lejocelyn@host86-173-4-249.range86-173.btcentralplus.com) |
01:13.37 | *** join/#gllug lejocelyn_ (~lejocelyn@host86-173-4-249.range86-173.btcentralplus.com) |
02:43.31 | *** join/#gllug DiscordianUK (~ch@fedora/DiscordianUK) |
03:10.42 | *** join/#gllug Leeds (~richardc@www.scorefive.com) |
05:40.07 | *** join/#gllug LeedsHK (~richardc@n219078059056.netvigator.com) |
05:43.43 | Leeds | oh look, there's two of me |
05:43.49 | Leeds | LeedsHK: hello Irene! |
05:44.19 | LeedsHK | :D |
05:44.55 | Leeds | feeling any better? |
05:45.03 | LeedsHK | yes |
05:45.07 | Leeds | good |
05:45.12 | LeedsHK | are you coming home? |
05:45.27 | Leeds | as I said, 3ish |
05:45.49 | LeedsHK | ok la |
05:46.11 | Leeds | now stop pretending to be me and go to bed :P |
05:46.58 | LeedsHK | i can close this chat? |
05:47.04 | Leeds | sure |
05:47.12 | LeedsHK | see ya later |
05:47.15 | Leeds | seeya |
07:34.10 | *** join/#gllug stafacus (~stafacus@78.146.208.237) |
07:54.59 | *** join/#gllug Leeds (~richardc@n219078059056.netvigator.com) |
08:19.00 | bilarh | random question of the day: anyone know anything about lathes and chuck mountings? |
08:34.13 | wethrin | a woodworker would |
08:39.33 | bilarh | is that like a wood chucker? :S |
08:51.09 | wethrin | a woodchuck chucks would |
08:51.11 | wethrin | wood |
09:14.44 | gregj | chuck norris knows |
09:15.55 | bilarh | wethrin: nice save their, i was just about to rip into you |
09:23.53 | wethrin | i hadn't had coffee! |
09:26.47 | bilarh | i don't think i oculd get out of bed without coffee |
09:26.50 | bilarh | i'd be so constantly tired |
09:27.14 | bilarh | of course i realise that if i didn't drink so much coffee, i'd have to go to bed earlier and thus not be as tired |
09:27.20 | bilarh | but, i chose to quietly ignore that :P |
09:27.34 | bilarh | and either way, coffee tastes nice |
09:27.35 | wethrin | I'm waiting for coffee to cool so I can drink it |
09:27.52 | bilarh | concludes his mini rant |
09:36.42 | *** join/#gllug zplinux (~zplinux@213.8.57.217) |
09:36.47 | zplinux | hi |
09:36.52 | zplinux | hali: ping |
09:37.14 | zplinux | can anyone exaplian this simple chore |
09:37.36 | zplinux | how to combine mysql two hosts set-up as amster master with IP alias |
09:37.48 | zplinux | so one IP access the mysql server |
09:37.59 | bilarh | that doesn't sound all that simple at all :P |
09:38.23 | zplinux | what address should I bind each server to? |
09:39.17 | zplinux | say host1 is 1.1.1.1 and host2 is 1.1.1.2 |
09:39.25 | zplinux | IP alias is 1.1.1.3 |
09:39.39 | zplinux | what address should each mysql bind to? |
09:40.16 | wethrin | why do you want a two master setup, rather than a master-slave setup? |
09:40.50 | zplinux | hi wethrin |
09:41.03 | zplinux | I should point out that I am new to mysql |
09:41.44 | wethrin | Yes. And a two master setup is far harder than a master-slave setup |
09:41.47 | zplinux | when all is ok, host1 servers the mysql. when it dies hosts2 serves it |
09:42.15 | zplinux | in case host1 returns to life, it regain the master role |
09:42.48 | bilarh | that's a master-slave set-up not a master-master |
09:42.49 | zplinux | so I dont need to bother with restoring th databases from the slave to the master |
09:43.43 | zplinux | clients will access one IP address and not bother with which server is actually serving it |
09:44.03 | zplinux | master master allows the first master to regian control - right? |
09:44.09 | zplinux | regain |
09:44.26 | zplinux | bilarh: am I wrong? |
09:44.43 | bilarh | well, i'm not actually that accustomed to mysql myself |
09:44.44 | bilarh | but |
09:44.59 | bilarh | even oracle doesn't do the "one ip access point" in rac, does it |
09:45.09 | bilarh | that part is handled by the client |
09:45.13 | bilarh | but either way |
09:45.25 | zplinux | how ? |
09:45.32 | bilarh | you could probably bodge it up with mysql running in master-slave mode |
09:45.44 | bilarh | and with heartbeat (for example) to handle the ip-alias-failover |
09:45.59 | zplinux | should mysql access from client go though a script routing calls to each server? |
09:46.19 | bilarh | with rac, afaik the client picks a server that is available |
09:46.31 | zplinux | yes - I intend to use a IP alias - heartbeat setup |
09:46.40 | bilarh | but it's all handled by the libraries that rac-clients are linked against |
09:46.53 | zplinux | mysql? |
09:47.20 | bilarh | no, rac |
09:47.44 | zplinux | I ment, any clue to how to achive this in mysql? |
09:47.51 | zplinux | ~seen Leeds |
09:47.53 | ibot | leeds is currently on #gllug (1h 52m 54s), last said: 'seeya'. |
09:47.54 | bilarh | no :) |
09:48.00 | bilarh | http://dev.mysql.com/doc/refman/5.1/en/replication-implementation.html |
09:48.12 | zplinux | let me read this |
09:48.33 | zplinux | where is hali , when I need him |
09:48.39 | zplinux | isnt he the db guy? |
09:49.17 | bilarh | amongst other things |
09:49.18 | zplinux | by the way workign on weekends aint that bad |
09:49.29 | bilarh | rather you than me |
09:49.35 | zplinux | :-) |
09:49.45 | zplinux | 14 days to the tour de france - right? |
09:50.01 | zplinux | so I must finish the project asap |
09:51.13 | wethrin | working on weekends is fine, if you get two days offin the week |
09:51.31 | bilarh | wethrin: yes, unfortuantely that is very rare nowadays it seems |
09:51.37 | bilarh | i don't even get overtime anymore |
09:51.41 | bilarh | only time actually worked |
09:51.52 | bilarh | not even travel time, which sucks seeing as i live an hour away from work |
09:51.54 | wethrin | so that's about 5 hours a week for you? |
09:51.54 | wethrin | :) |
09:52.03 | bilarh | yeah hehe |
09:52.08 | bilarh | but i mean weekend work :P |
10:21.41 | zplinux | I am falling a sleep reading about keepalived |
10:21.45 | zplinux | err |
10:25.10 | antiphase | MySQL is bollox, so having a redundant functioning master setup which can fail over without intervention is unlikely |
10:45.21 | hali | you'd have to go oracle for a fully active/active single connection string config |
10:45.49 | hali | for mysql i'd say use a single master with failover and have select slaves if required |
10:45.52 | zplinux | hali:!! |
10:46.24 | zplinux | can you please explain that to me? |
10:46.44 | zplinux | " have select slaves if required" - what does that mean? |
10:46.56 | hali | replicated read-only copies of the db |
10:47.34 | zplinux | so what happens when the master dies? |
10:48.17 | antiphase | Someone phones you up and you fix it |
10:48.30 | zplinux | this is bad |
10:48.41 | antiphase | You just have to make sure that doesn't happen :P |
10:48.59 | zplinux | system should survive this |
10:49.10 | antiphase | Reliable power, plus redundant system components, plus resilient storage, plus backups |
10:49.16 | hali | multiple ways you could do that, cleanest way would be to have shared storage and something like vcs |
10:49.22 | bilarh | s/doesn't happen/you are not on call when it happens/ |
10:49.24 | zplinux | I want to watch the tour de france and not worry about that |
10:49.41 | antiphase | If it was never going to happen, you'd never be on call. Shit happens ;) |
10:49.50 | zplinux | hali: when are you usually here? |
10:50.15 | bilarh | antiphase: that's not very egoistic thinking of you... clearly you can't have worked for big enough companies yet :P |
10:50.18 | zplinux | anti: I am the only one that can fix this - so I am allways on call |
10:50.28 | antiphase | Sucks to be you |
10:50.54 | zplinux | what? |
10:51.03 | antiphase | You also haven't cited the failure modes which aren't due to a hardware failure, like when someone goes DELETE FROM table; |
10:51.04 | zplinux | bilarh: ha? |
10:51.06 | antiphase | oops |
10:51.26 | zplinux | ok, ok sow down |
10:51.31 | zplinux | slow down |
10:51.36 | zplinux | let me explain. |
10:51.42 | antiphase | I would assert that hardware failure is likely to be the least of your problems when running a database |
10:51.55 | antiphase | doesn't do databases for this reason |
10:51.56 | zplinux | no one will ever have direct access to the mysql but the programs I setup |
10:52.24 | antiphase | Like children and animals on television, I never approach storage or databases except when someone else looks after them |
10:52.27 | zplinux | anti: I also dont like db |
10:52.43 | zplinux | yet I do admire ppl writing db |
10:52.56 | bilarh | antiphase: does that mean you *do* approach children when you're not being filmed? ;) |
10:53.00 | zplinux | back to the subject |
10:53.31 | antiphase | I admire fuck all, because I am a sysadmin. IT is all hateful software and hardware that goes wrong |
10:53.33 | zplinux | I am trying to understand what options I got in the free-software community to allow |
10:53.53 | zplinux | the system to sustain a disaster |
10:54.15 | antiphase | What sort of disaster? Direct nuclear strike or someone tripped over the power cable? |
10:54.38 | zplinux | second one - I guess |
10:54.40 | bilarh | zplinux: look, why don't you just have a master-slave setup as described on that mysql page and add an ip-alias-failover with heartbeat... then if something goes wrong, accept that your new master will have to become a slave of the "slave" so that the "slave" becomes the master... what's up with that? |
10:55.19 | zplinux | bilarh: excelent! |
10:55.21 | bilarh | human factor is a much, much bigger risk than hardware in my experience... especially the "bilarh" factor :P |
10:55.31 | zplinux | I am just tring to understand this |
10:55.46 | antiphase | Try stuff first, then ask if it doesn't work |
10:55.48 | bilarh | but you need to have manual intervention to get it working.. you can't automatically get it to work |
10:55.58 | zplinux | I am allready going to apply heartbeat and IP aliasing |
10:56.17 | bilarh | ok good, i'd also strongly recommend you get one of those APC power switch things |
10:56.28 | bilarh | or even two (one for each server) if you're paranoid |
10:56.46 | zplinux | bilarh: so to recover , in case the master dies |
10:56.57 | bilarh | then the slave will take over |
10:57.02 | zplinux | I will need to dump the slave db, restore them to the master and restart? |
10:57.06 | bilarh | but then before you bring the master back online, you have to make the slave the new master |
10:57.10 | bilarh | and the old master has to become the slave |
10:57.14 | bilarh | no there's no dumping |
10:57.25 | bilarh | you just designate your slave to become a master |
10:58.00 | bilarh | http://dev.mysql.com/doc/refman/5.1/en/replication-solutions-switch.html figure 16.5 |
10:58.22 | zplinux | thanks , bilarh |
10:58.28 | zplinux | I am off to lunch now, |
10:58.38 | zplinux | printed a lot a reading matirial |
10:58.51 | bilarh | np |
10:58.52 | zplinux | will be back in case I got more questions |
11:49.05 | *** join/#gllug dick_turpin (~dick_turp@host217-34-163-30.in-addr.btopenworld.com) |
11:49.45 | dick_turpin | Afternoon all |
11:56.30 | bilarh | afternoon |
11:57.27 | dick_turpin | bilarh: How we doing? |
11:58.02 | bilarh | you royal now |
11:58.12 | bilarh | U SLAAAAG etc |
12:01.03 | dick_turpin | Huh I've got nowhere with mailing to the list both party's have sent me URL's to suppose evidence of 'their' innocence I am left in a position where I can get the mails but cant send which interestingly means I cant unsubscribe either! |
12:03.19 | dick_turpin | supposed |
12:24.15 | *** join/#gllug stu_ (~stu@dyn1247-66.vpn.ic.ac.uk) |
12:43.30 | *** join/#gllug DiscordianUK (~ch@fedora/DiscordianUK) |
12:51.22 | *** part/#gllug dick_turpin (~dick_turp@host217-34-163-30.in-addr.btopenworld.com) |
13:52.14 | *** join/#gllug dick_turpin (~dick_turp@host217-34-163-30.in-addr.btopenworld.com) |
13:52.24 | *** join/#gllug stu_ (~stu@94-30-86-45.xdsl.murphx.net) |
13:58.00 | *** join/#gllug stu_ (~stu@dyn1241-59.vpn.ic.ac.uk) |
14:02.06 | *** join/#gllug celesteh (~celesteh@sblug/member/celesteh) |
14:28.16 | dick_turpin | wethrin: Can you have a look see if a mail from me is in moderation folder on the list |
14:29.57 | wethrin | nothing pending |
14:30.20 | dick_turpin | hm |
14:31.37 | dick_turpin | My ISP now says they've fixed it, replied against your 'testin' post. The mail went and no bounce back but its not showing on the list (Blackhole?) |
14:31.48 | dick_turpin | testing |
14:46.35 | mindbndr | z00dax: any news for the centos rss ? :) |
15:02.25 | *** join/#gllug Armand (~me@mail.justit.co.uk) |
15:17.33 | dick_turpin | wethrin: Aha! Email has been semi accepted sitting in greylist apparently now waiting for my ISP's server to try again |
15:36.14 | *** part/#gllug dick_turpin (~dick_turp@host217-34-163-30.in-addr.btopenworld.com) |
17:30.39 | *** join/#gllug DiscordianUK (~ch@fedora/DiscordianUK) |
21:17.56 | *** join/#gllug Armand (~me@host86-183-128-14.range86-183.btcentralplus.com) |
22:53.35 | *** join/#gllug DiscordianUK (~ch@fedora/DiscordianUK) |
23:09.22 | *** join/#gllug Leeds (~richardc@n219078059056.netvigator.com) |