00:18.34 | *** join/#utos mindjuju (~mindjuju@173-10-26-193-BusName-utah.ut.hfc.comcastbusiness.net) |
01:43.35 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
04:51.11 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
07:56.11 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
08:05.03 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
08:42.51 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
15:15.20 | *** join/#utos fugue88 (~dsowen@173-165-137-65-utah.hfc.comcastbusiness.net) |
15:36.18 | *** join/#utos mindjuju (~mindjuju@173-10-26-193-BusName-utah.ut.hfc.comcastbusiness.net) |
15:38.28 | *** join/#utos mindjuju (~mindjuju@173-10-26-193-BusName-utah.ut.hfc.comcastbusiness.net) |
16:55.36 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
17:51.41 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
19:09.51 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
19:26.04 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
19:35.37 | *** join/#utos josePHPagoda (~josephnex@71-209-43-153.bois.qwest.net) |
21:11.04 | *** join/#utos tristanbob_l (~tristan@oalug/member/tristanbob) |
21:14.06 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
21:19.00 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
21:22.03 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
21:38.08 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
21:40.24 | *** join/#utos bgerber (~gerberb@suse103.zenez.com) |
22:45.43 | *** join/#utos josePHPagoda (~josephnex@24-116-177-210.cpe.cableone.net) |
22:56.15 | josePHPagoda | hi guys! |
22:56.52 | josePHPagoda | so i'm wondering, when I'm creating a tarball, if I do a --atime-preserve, when I extract that file, will it have the same mtime on all of the destinations that I extract it to? |
22:58.31 | herlo | no |
22:58.54 | herlo | josePHPagoda: that will preserve the latest access time, you probably want to preserve the modified time |
22:59.40 | fugue88 | In fact, it preserves atime on the original file; the flag doesn't change what is recorded in the tar (it probably always records the original atime or something). |
22:59.54 | josePHPagoda | yeah |
23:00.02 | josePHPagoda | we're setting up our etags stuff |
23:00.15 | herlo | josePHPagoda: look at --mtime |
23:00.31 | fugue88 | josePHPagoda: Your webservers transfer everything as tars? ;P |
23:00.34 | josePHPagoda | wanna make sure it's time is consistent between servers |
23:00.39 | josePHPagoda | we do use tars for a lot |
23:00.41 | josePHPagoda | :-P |
23:00.47 | josePHPagoda | it's a simple tool that works quickly |
23:00.50 | josePHPagoda | :-D |
23:00.55 | herlo | josePHPagoda: I heard from Heartsbane last night that you guys don't use a standard logger |
23:01.06 | josePHPagoda | that is correct |
23:01.09 | fugue88 | Standard loggers are simple tools that work quickly. |
23:01.15 | fugue88 | :D |
23:01.30 | josePHPagoda | they do, but they don't aggregate and insert records the way we want |
23:01.32 | josePHPagoda | :-P |
23:01.35 | herlo | josePHPagoda: I suggest looking at rsyslog if you need to store your logs other than in the filesystem and start using a standard logging |
23:01.40 | herlo | tool |
23:01.48 | josePHPagoda | we're using an analytics db |
23:01.52 | josePHPagoda | and dumping them directly into there |
23:01.55 | josePHPagoda | :-D |
23:01.56 | herlo | josePHPagoda: rsyslog can use a db |
23:02.01 | josePHPagoda | not monet |
23:02.01 | herlo | and log directly there |
23:02.03 | josePHPagoda | sadly :-( |
23:02.06 | fugue88 | Use 0mq. It's cooler. |
23:02.08 | josePHPagoda | though we are using syslog packets |
23:02.08 | herlo | josePHPagoda: money is dome |
23:02.14 | herlo | s/dome/dumb/ |
23:02.22 | herlo | gah! |
23:02.22 | josePHPagoda | monet or money? |
23:02.26 | josePHPagoda | :-P |
23:02.30 | herlo | repeats what he meant |
23:02.33 | herlo | monet is dumb! |
23:02.37 | josePHPagoda | lol |
23:02.40 | herlo | kicks his fingers |
23:02.44 | josePHPagoda | why do you say that? |
23:02.55 | fugue88 | monet is dead |
23:03.01 | fugue88 | He can't talk. |
23:03.05 | herlo | indeed |
23:03.07 | josePHPagoda | lol |
23:03.13 | josePHPagoda | have you guys used monetdb? |
23:03.22 | josePHPagoda | (and they've got commits happening fairly often, just saying) |
23:03.25 | josePHPagoda | :-P |
23:03.29 | fugue88 | So, they're unstable? |
23:03.30 | herlo | josePHPagoda: seriously though, rsyslog could work with whatever, it's just an odbc connection, right? |
23:03.47 | josePHPagoda | but we're open to suggestions for a better datastore if you guys know of any |
23:03.50 | josePHPagoda | :-P |
23:03.56 | josePHPagoda | rsyslog is good, but won't catch everything sadly |
23:03.59 | herlo | josePHPagoda: postgresql |
23:04.03 | herlo | josePHPagoda: how so? |
23:04.10 | herlo | rsyslog won't catch what? |
23:04.16 | josePHPagoda | custom apps unfortunately |
23:04.19 | josePHPagoda | :-( |
23:04.20 | herlo | it's just a logger, your app sends it to the logger |
23:04.24 | josePHPagoda | yeah |
23:04.37 | josePHPagoda | we're writing something that covers people that use file based logs, and not stream based logs |
23:04.41 | herlo | you mean custom *poorly* *written* apps? |
23:04.45 | josePHPagoda | yeah |
23:04.49 | fugue88 | heh |
23:04.51 | josePHPagoda | sadly, we need to accomodate those too |
23:04.57 | fugue88 | Now you have *two* problems. |
23:05.00 | herlo | josePHPagoda: how does that matter? |
23:05.12 | herlo | file based or stream based, rsyslog can do both at the same time |
23:05.18 | herlo | isn't kidding |
23:05.30 | fugue88 | herlo: As logging sources, or just as logging destinations? |
23:05.46 | herlo | destination |
23:06.09 | herlo | sources are apps, I suppose they could be files if there were some way to process them into a standard logger |
23:06.16 | fugue88 | josePHPagoda: Simple workaround for apps that insist on logging to files: tail -F redir'd to syslog. |
23:06.26 | herlo | fugue88: better is a named pipe |
23:06.32 | fugue88 | Oooh, yeah! |
23:06.34 | josePHPagoda | there are a few interesting issues |
23:06.37 | herlo | rsyslog can pull from that |
23:06.50 | josePHPagoda | that logger stuff will be coming shortly, and we may reevaluate it, i dunno |
23:06.51 | herlo | pretty easily in fact |
23:07.02 | herlo | josePHPagoda: I highly recommend looking back at rsyslog |
23:07.04 | josePHPagoda | rsyslog isn't bad by any means, we just had trouble getting it to catch a few things |
23:07.16 | herlo | like what |
23:07.17 | josePHPagoda | but maybe when we revisit it more thoroughly, we'll catch what we missed |
23:07.33 | josePHPagoda | hang on though, wanna get this tar thing working right |
23:07.47 | herlo | hasn't seen anything solid that rsyslog couldn't do from your statement |
23:07.50 | fugue88 | You like a little sunk-cost, fallacy, eh? |
23:07.58 | fugue88 | josePHPagoda: ^ |
23:08.02 | josePHPagoda | if I do a --mtime then, when the tarball is extracted, it'll have a consistent time? |
23:08.17 | herlo | josePHPagoda: not sure off hand, test it yerself |
23:08.24 | herlo | use touch if you need to modify the mtime |
23:08.39 | herlo | will be back in a bit, gotta go replace my pants |
23:08.44 | fugue88 | tmi |
23:08.51 | herlo | fugue88: not really |
23:08.56 | fugue88 | :) |
23:09.13 | herlo | I bought some carpenter jeans and last night they ripped across the rear pocket |
23:09.21 | herlo | It's really the only pair I like |
23:09.26 | fugue88 | :( |
23:09.28 | herlo | so I am going to go buy some more to replacfe them :) |
23:09.32 | herlo | s/fe/e/ |
23:10.05 | josePHPagoda | fugue88, herlo, it isn't that, it's that i'm wanting to get this tar thing worked out first... our main issues that we had with rsyslog had to do with aggregating the information, but since we got monet to use odbc, that might not be so much of an issue anymore, i dont know, like I said, we'll have to revisit it... trust me, if we lived based on sunk costs, we'd be horribly broken right now... we've refactored many many times |
23:10.08 | josePHPagoda | :-P |
23:10.27 | fugue88 | :D |
23:12.36 | josePHPagoda | we'll just have to revisit our list of requirements and see what it says |
23:15.56 | fugue88 | josePHPagoda: I think tar will try to set the archived mtime on extraction by default. |
23:16.31 | josePHPagoda | that's what we're hoping for |
23:16.34 | josePHPagoda | just running some tests |
23:16.37 | josePHPagoda | :-D |
23:17.21 | fugue88 | josePHPagoda: Yep, preserved mtime, but not atime or ctime (mtime is usually what you care about though). |
23:17.33 | josePHPagoda | that is exactly what we care about |
23:17.35 | josePHPagoda | :-D |
23:19.30 | josePHPagoda | just fyi guys, regardless of logging solution we use, we'll be letting people specify a remote syslog server if they want logs streamed to them |
23:19.33 | josePHPagoda | :-D |
23:46.48 | josePHPagoda | yay! |
23:46.51 | josePHPagoda | it worked |
23:46.53 | josePHPagoda | we're not using etags properly |
23:47.00 | josePHPagoda | err |
23:47.03 | josePHPagoda | we're now using etags properly |
23:47.06 | josePHPagoda | :-P |
23:47.10 | josePHPagoda | talk about a typo changing the meaning |