[02:35:47] New patchset: Tim Starling; "1/100 sampling for banner impressions" [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6779 [02:36:05] New review: gerrit2; "Lint check passed." [operations/puppet] (production); V: 1 - https://gerrit.wikimedia.org/r/6779 [02:47:45] New review: Tim Starling; "(no comment)" [operations/puppet] (production); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6779 [02:47:48] Change merged: Tim Starling; [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6779 [03:28:25] New patchset: Asher; "don't try to cache large media objects in the frontend instance set stream buffer 10M in frontend enable streaming from the backend for objects > 64M" [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6780 [03:28:43] New review: gerrit2; "Lint check passed." [operations/puppet] (production); V: 1 - https://gerrit.wikimedia.org/r/6780 [03:29:27] New review: Asher; "(no comment)" [operations/puppet] (production); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6780 [03:29:30] Change merged: Asher; [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6780 [03:34:21] New patchset: Asher; "beresp.stream_pass_bufsize isn't actually in varnish 3.0.2" [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6781 [03:34:39] New review: gerrit2; "Lint check passed." [operations/puppet] (production); V: 1 - https://gerrit.wikimedia.org/r/6781 [03:34:43] New review: Asher; "(no comment)" [operations/puppet] (production); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6781 [03:34:45] Change merged: Asher; [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6781 [03:37:07] New patchset: Asher; "dash vs. underscore. underscore wins." [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6782 [03:37:25] New review: gerrit2; "Lint check passed." [operations/puppet] (production); V: 1 - https://gerrit.wikimedia.org/r/6782 [03:37:28] New review: Asher; "(no comment)" [operations/puppet] (production); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6782 [03:37:30] Change merged: Asher; [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6782 [07:20:02] !log power cycled db45 (crashed dewiki slave) [07:20:06] Logged the message, Master [08:00:59] !log upgrading/rebooting the last couple sq* servers [08:01:02] Logged the message, Master [08:39:14] New patchset: Dzahn; "minor fixes to language and license columns" [operations/debs/wikistats] (master) - https://gerrit.wikimedia.org/r/6787 [08:39:15] New patchset: Dzahn; "enhance siteinfo() fetching - debug error codes" [operations/debs/wikistats] (master) - https://gerrit.wikimedia.org/r/6788 [08:39:15] New patchset: Dzahn; "needed to handle wikis with API siteinfo but not API stats, fix sorting by http" [operations/debs/wikistats] (master) - https://gerrit.wikimedia.org/r/6789 [08:40:13] New review: Dzahn; "(no comment)" [operations/debs/wikistats] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6787 [08:40:15] Change merged: Dzahn; [operations/debs/wikistats] (master) - https://gerrit.wikimedia.org/r/6787 [08:41:35] New review: Dzahn; "(no comment)" [operations/debs/wikistats] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6788 [08:41:37] Change merged: Dzahn; [operations/debs/wikistats] (master) - https://gerrit.wikimedia.org/r/6788 [08:42:36] New patchset: Dzahn; "needed to handle wikis with API siteinfo but not API stats, fix sorting by http, fix red gerrit marks" [operations/debs/wikistats] (master) - https://gerrit.wikimedia.org/r/6789 [08:43:17] New review: Dzahn; "(no comment)" [operations/debs/wikistats] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6789 [08:43:19] Change merged: Dzahn; [operations/debs/wikistats] (master) - https://gerrit.wikimedia.org/r/6789 [09:27:02] !log rebooting bits varnish sq68-70 one by one.. [09:27:05] Logged the message, Master [10:42:00] New patchset: Dzahn; "adding interactive server upgrade-helper script" [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6791 [10:42:18] New review: gerrit2; "Lint check passed." [operations/puppet] (production); V: 1 - https://gerrit.wikimedia.org/r/6791 [10:43:26] New patchset: Dzahn; "adding interactive server upgrade-helper script" [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6791 [10:43:44] New review: gerrit2; "Lint check passed." [operations/puppet] (production); V: 1 - https://gerrit.wikimedia.org/r/6791 [10:44:51] New review: Dzahn; "just putting a helper script in misc/scripts/" [operations/puppet] (production); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6791 [10:44:53] Change merged: Dzahn; [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6791 [11:09:50] !log squids - sq* done. all latest kernel and 0 pending upgrades. [11:09:53] Logged the message, Master [11:10:42] * ^demon hands mutante a cookie :) [11:11:21] ty, demon;) [11:11:42] <^demon> yw. I mean, who says no to a cookie :) [11:12:31] just my browser, sometimes:) [11:13:49] the helper script gives you a kitten if it detects you are all done. make it more fun :p [11:18:13] !log continuing with upgrades/reboots in amssq* on the side during the day [11:18:16] Logged the message, Master [11:20:16] <^demon> mutante: `ack --thppt` [11:21:52] oh? hashar wants to make sure it is removed in favor of grep-ack? [11:24:49] <^demon> ack-grep, app::ack, betterthangrep, it's all the same thing :) [11:24:58] <^demon> A much much faster version designed for searching source code. [11:28:36] yea, adding classes for them is fine, it is the discussion what should be in base and what should not though that always has different opinions (and the stuff like ack vs. grep-ack). just like global vim config and stuff.. [11:29:45] <^demon> *nod* [11:30:23] <^demon> Useful tools are useful, but cluttering the base install with a bunch of tools -> $maintainability-- [11:30:43] and editors, re:Joe [11:31:05] <^demon> Joe was only because brion liked it, iirc ;-) [11:31:23] better joe than mc ;) [11:31:38] just saying i can imagine the next editor request [11:32:39] <^demon> "provide me with a google docs bridge so I can edit site configuration from my browser" [11:32:42] <^demon> ;-) [11:32:51] i dont hate joe, i think it's ok to offer labs users one vim alternative..but maybe not 10 [11:33:29] hehe @ google docs config ,,yea [11:34:22] thats why i'd like to see private etherpads , hah [11:41:49] New patchset: Dzahn; "minor fix and tabbing in upgrade-helper script" [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6792 [11:42:08] New review: gerrit2; "Lint check passed." [operations/puppet] (production); V: 1 - https://gerrit.wikimedia.org/r/6792 [11:42:31] New review: Dzahn; "(no comment)" [operations/puppet] (production); V: 1 C: 2; - https://gerrit.wikimedia.org/r/6792 [11:42:33] Change merged: Dzahn; [operations/puppet] (production) - https://gerrit.wikimedia.org/r/6792 [12:43:18] New review: Demon; "What was the reason for this again?" [operations/puppet] (production) C: 1; - https://gerrit.wikimedia.org/r/4083 [12:49:25] !log pushing out virtual host for wikimania2013 wiki. sync / apache-graceful/all [12:49:29] Logged the message, Master [12:51:35] Reedy: Thehelpfulone ^^ [12:52:09] ooh a ping [12:52:10] :D [12:52:32] Thehelpfulone: it's on the Apaches now [12:52:50] ok [12:53:17] you can go ahead with wiki install i guess, now brings you to that start page [12:53:27] instead of wikimediafoundation.org [12:53:45] the wiki's been installed partly already I think mutante [12:53:54] there was an email to that newprojects mailing list [12:54:07] http://wikimania2013.wikimedia.org/wiki/Main_Page it just needs all the relevant settings etc copied [12:54:12] ok, shift+reload and see the difference [12:54:16] ok [12:56:04] do you know which ones those are https://bugzilla.wikimedia.org/show_bug.cgi?id=36477 is the bug - see my first comment. [12:57:35] Thehelpfulone: updated BZ, but no i don't know which exact settings you need, Reedy will for sure [12:58:28] yep sure no problem, casey's comments were "This wiki should have the same settings as the other Wikimania wikis for the logo, project name, extra namespaces, ForceUIMsgasContentMessage, anon editing restrictions, etc. (previous bugs: bug 18740, bug 13547)" [13:00:59] that can wait for Reedy though who should be in the office in a few hours [13:01:16] mutante, is the initial crat made through the database? [13:01:47] <^demon> addwiki.php doesn't make an initial crat, no. [13:02:23] Thehelpfulone: ^demon knows better than i do about anything after DNS and Apache config [13:02:28] heh [13:02:47] <^demon> DNS? I know zilch. [13:02:57] ^demon: yeah I meant on previous requests a user was made a crat to be able to assign rights etc [13:02:59] usually that is like the point of giving it from ops to devs/devops [13:03:13] <^demon> Thehelpfulone: Perhaps someone edited the database and made them a crat? [13:03:20] yeah that's what I said :P [13:03:45] so who do I need to ask to be made the initial crat, I'm taking that role of handing out rights etc [13:04:08] <^demon> Any +shell user can, but since Reedy made the wiki I'd ask him. [13:04:55] ok [13:16:07] hey apergos, are you around? [13:19:21] yes [13:19:43] sorry, I was afk redesigning some code that I had written in a completely braindead way [13:19:49] np [13:19:54] (let me grab ottomata) [13:19:54] what's up? [13:20:22] i'm here [13:20:43] ottoman and i are curious to hear what happened with the fund-raising filter yesterday [13:20:52] and what actions we can take to make our lives easier in the future [13:21:05] and is storage3 an additional filter box? [13:21:41] I'm absolutely the wrong prson to ask. I can tell you what I saw and what I did but the prson who knows about this is (I think) jefff and he's on ... ny? sf? time [13:21:54] so the locke logs are copied to storage3 periodically [13:22:18] the partition where they were to be copied was not accessible, spewing a bunch of raid related messages in the log [13:22:21] so copies failed [13:22:37] apergos, how are they copied? [13:22:42] cron on storage3? [13:22:50] so locke started getting full-ish on that one partition with the logs and someone using that data saw that it wasn't available [13:22:55] uh huh a cron job on storage3 [13:22:58] k [13:23:11] and they are deleted on locke by the same cron? [13:23:24] yes [13:23:35] from /a/squid/archive? [13:23:35] of course we don't delete if the copy doesn't happen [13:23:55] you will hate me but I already don't remember the directory [13:23:59] it's in the log maybe [13:24:04] (sysadmin log) [13:24:09] yeah, i think that is it [13:24:20] it looks full and there is a logrotate file putting stuff there [13:24:24] ottomata: i think you need to apply for a storage3 account :) [13:24:24] so then I said I would try to reboot storage and see if we got the partitions back, [13:24:37] since it was already broken as things were [13:24:49] and on reboot those partitions didn't come up [13:25:02] so I skipped the moount for the two raid partitions [13:25:18] and that's where storage3 is, out of action as far as log storage. [13:25:31] at that point I manually copied the logs from locke to hume, [13:25:37] gzipped them there, [13:25:49] what's hume? [13:25:56] (ah, after hupping udp2log) [13:26:04] and then removed the copies from locke. [13:26:15] did you copy from the live log files then? [13:26:20] and not the archived ones? [13:26:27] (if you needed to hup) [13:26:30] hume is a host we use for .. well it has had copies of logs in the past but typically it is for more computationally intensive tasks by developers, long running scripts and such [13:26:38] ok [13:27:05] no. I moved the logs to a different directory, hupped the upd2log, thn copied those logs which were not longer being updated to hume, then gzipped them, then removed them from locke [13:27:32] this is ordinarily what the cron script does [13:27:52] I simply did it manually, with the exception that instead of copying to storage3 I put them in a directory on hume (see the sysadminlog) [13:28:12] all clear? [13:28:16] really? [13:28:23] what? [13:28:26] it doesn't copy the already zipped logs from /a/squid/archive? [13:28:34] there is a logrotate script that is moving files there [13:28:35] sorry [13:28:39] I read the cron job and did what it does [13:28:41] / var/log/archive [13:28:42] hm [13:29:04] this is for bannerimpressions and one other log which (again) I already forget what it is [13:29:19] I copied a total of 4 files, two of each type. [13:29:21] ah no, /a/squid/archive [13:29:21] is that [13:29:25] hm [13:29:38] ah they have their own directory [13:29:55] /a/squid/fundraising [13:30:02] quick irrelevant Q [13:30:03] obviously the ones now on hume should *not* be removed [13:30:09] why are all these logs in a directory called 'squid' [13:30:13] the logs are not all from squid [13:30:14] we need to keep them until jeff can get things sorted [13:30:27] do not look at me, thanks :-P [13:31:11] haha [13:31:28] i have such a strong urge to organize this whole system [13:31:34] but i'd probably break lots of things doing it [13:31:34] oh sorry it runs on locke [13:31:37] not on storage3 [13:31:40] the cron? [13:31:43] as who? [13:31:46] rotate_logs_and_copy_to_storage3.pl here's the sciptname [13:32:00] file_mover I guess [13:32:23] the only way for me to be sure is go look everything up again [13:32:59] yeah found it [13:33:01] file_mover [13:33:30] found this [13:33:30] http://wikitech.wikimedia.org/view/Fundraising_Analytics/Impression_Stats [13:33:36] great [13:34:18] I didn't disable the cron job, it runs and fails right now [13:34:32] a bit spammy but presumably it will get sorted out later today [13:35:30] so now you know as much as (or more than) me [13:36:10] ok cool [13:36:17] thanks apergos [13:36:24] yw [13:36:26] yeah thanks! [13:36:28] any idea where this whole /a thing came from? [13:36:36] i had thought it was erik for 'analytics' [13:36:38] but i guess not [13:36:39] all that stuff well predates me [13:36:46] years and years old. [13:37:11] could be /a for apache. or /a because we chose a letter at random for all I know [13:50:16] so, apergos, who are you waiting on to fix hume/storage3 stuff? [13:50:28] I hope jeff will look at it [13:50:32] iirc it's his "baby" [13:50:37] what's his sn? [13:50:47] nick? [13:51:02] he's not on right now, wrong timezone [13:51:42] aye, just curious what it is [13:51:47] he has had issues with storage3 in the past so I bet he has some fallback plans up his sleeve [13:52:37] aye ok [13:52:50] including adding monitoring for disk space there :p ? [13:53:23] Jeff_Green is the nick [13:53:36] it's not a matter of disk space [13:53:44] it is afaict a hardware issue [13:56:47] oh hm, ok [13:57:28] you did look at the sysadmin log right? :-P [14:01:35] how do I do that? [14:01:44] the nagios notices? [14:03:38] RT-2907: rsync: change_dir "/archive/udplogs" failed: No such file or directory (2) [14:03:41] there he is:) [14:04:05] nice timing [14:04:43] ugh, storage3? wtf [14:05:00] there were backup failures [14:05:09] rsync -ar --delete /archive/udplogs/ file_mover@hume.wikimedia.org: [14:05:13] /archive/udplogs/ [14:05:19] storage3 is fubar [14:05:22] and rsync -ar /archive/jenkins_builds/ logmover@storage3.pmtpa.wmnet: [14:05:33] ottomata just asked about it before you joined [14:05:41] put it in an RT for now [14:05:48] root@storage3:~# mount /a [14:05:48] mount: special device UUID=cc471f5e-062d-4d73-83a2-8946667a13e5 does not exist [14:05:59] huh [14:06:07] yeah, [14:06:12] see the sys admin log [14:06:28] looking [14:06:35] where is this sysadmin log? [14:06:44] http://wikitech.wikimedia.org/view/Server_admin_log [14:07:02] i am crying that this was not a pageable event [14:07:05] ah and those are teh !log messages people type here in here? [14:07:10] yes they are [14:07:23] well I don't think I broke anything [14:07:36] nothing filled up before I got called in [14:08:11] i don't think the issue is full disk [14:08:14] no [14:08:19] i think it's, as usual, RAID [14:08:21] yes [14:08:37] megacli like I say was just returning 0 and no contents [14:08:56] I didn't trry to dick around in any radi bioses or whatever [14:09:00] has rob h been alerted? [14:09:12] I doubt it [14:09:25] afaik dicking around with RAID bioses remotely is problematic anyway [14:09:34] at least on these machines [14:09:35] well I could have looked at it [14:10:22] but seemed like I wouldn't have been able to d much beyond look [14:10:58] I'm going to delete this ticket daniel posted and start with the RAID failure [14:11:07] ok. I didn't see the ticket [14:11:22] he entered one for the backups failures, but that's just a symptom [14:11:27] you have the useful info at this point [14:11:34] eh yeah, i just made that earlier when i read email [14:11:48] before i saw you talking about further issues [14:11:51] yeah [14:11:58] new one going in now [14:12:01] so right now storage3 is up without /a /archive mounted cause not detected as ready [14:12:30] yeah [14:12:35] all kinds of hell breaks loose in this case [14:12:38] possibly :-P [14:12:50] i'm not sure I accounted for this in the backup scripts [14:12:52] well no rsyncs go of course [14:13:34] as long as the ones that "rsync -var --delete /blah offhost:/blah" fail rather than purging everything on offhost [14:13:36] all the rsyncs are into lower level dirs I guess [14:14:04] yeah [14:14:07] that's good! [14:14:14] rob was on a little while ago, but dunno if on-site (or chris) [14:14:18] /home/file_mover/scripts/rotate_logs_and_copy_to_storage3.pl died: rsync: change_dir "/archive/incoming_udplogs" failed: No such file or directory (2) [14:14:26] this seemed to be the desired result [14:14:28] yep--that's good [14:15:45] do you guys know if nagios monitors disk space on all paritions by default? [14:15:54] on any puppetized machine? [14:16:02] the earlier failures are interesting (in an academic sort of way, not that interesting for fixing the real life issue) [14:16:28] [ 5.341656] megasas: Waiting for FW to come to ready state [14:16:28] [ 5.381673] megasas: FW in FAULT state!! [14:16:36] ain't it grand [14:16:39] apergos: which? [14:16:49] the earlier cron failures I mean, before my reboot [14:16:57] oh. looking [14:17:27] '/archive/udplogs' and '/archive/old_udplogs/' are identical (not copied) at /usr/local/bin/impression_log_rotator line 42 [14:17:27] impression_log_rotator died: Input/output error [14:17:38] yeah [14:17:51] whatever, we don't care right now but it is curious [14:17:53] it looks like all the RAID fell completely off line before you rebooted [14:17:59] it seemed to be off [14:18:10] that's why I rebooted, nothing was going to happen leaving it there [14:18:17] I couldn't do aany sort of diagnostics [14:18:31] identical may have been both dirs having no content [14:18:35] and a reboot had a possibility of clearing up some issue, or leaving things no worse off... [14:18:44] probably better, rlly [14:19:08] both empty, that's sure possible :-D [14:19:15] the 'identical' errors suggest the kernel wasn't in a healthy state about that partition [14:19:19] eh no [14:19:20] :-D [14:19:31] whereas after the reboot it's got the story straight :-( [14:19:42] yeah. no happy ending [14:20:26] !log stopped cron jobs on storage3 because of RAID failure [14:20:28] Logged the message, Master [14:20:42] I will take this opportunity to rant about RAID just once. I hate RAID. [14:20:59] really? why, what kind of RAID was it? [14:24:03] megacli, so some kind of LSI hardware RAID [14:24:17] why: because it's a false sense of security [14:24:39] RAID controllers are garbage, in my experience they have like a 20% failure rate [14:25:05] http://adminzen.org/backup/ :p [14:26:11] for storage though, do you really need a dedicated controller? [14:26:20] if you are just raiding for redundancy and not performance [14:26:37] md raid would be fine? [14:27:46] it'd be better imo [14:27:51] but this is also for mysql [14:28:51] ha, on a machine called storage3? [14:28:58] is it actually serving queries? [14:29:42] not really [14:29:53] then meh, md is fine too [14:30:09] the FR mysql architecure is db1008-->db1025/storage3 [14:30:33] it's an offsite slave, does dumps and other non-mysql backups [14:30:45] yeah md would probably be fine [14:36:00] yeah mysql was not running on the box when I got on it [14:36:10] yawp [14:36:23] (nor did I start it. uh uh.) [14:36:35] the long term plan is to carve out some space on the netapp [14:36:44] hmmmm [14:36:51] I guess people do that [14:37:54] it'd be better than relying on storage3 at least [14:38:20] and i don't want to double-book db1025 which is the only other option atm [14:40:12] Change abandoned: Hashar; "I am not sure why that was needed. Probably to manually setup branch bases documentation. That sur..." [operations/puppet] (production) - https://gerrit.wikimedia.org/r/4083 [14:40:40] agree [14:43:43] Jeff_Green: got me a question about something and your fingerprints are on it: [14:43:47] what is hume? [14:43:53] hahaha [14:44:12] notpeter: not sure! lemme check my notes [14:44:16] heh [14:44:19] "misc cmputing jobs that need a bit more cpu or memory, and we don't want them to run on fenari" [14:44:31] ah [14:44:35] mostly we don't want stuff like that on fenari, so that's where we put them [14:44:47] I got asked this very thing about an hour ago [14:45:12] notpeter: i think there was preexisting fundraising stuff on hume which I retooled a bit [14:45:31] gotcha gotcha [14:46:16] see, once you touch something it's yours forever, until the next person touches it [14:46:17] notpeter: which fingerprints where you referring to? [14:46:23] make sure to dump your notes on http://wikitech.wikimedia.org/view/Hume :-D [14:46:36] you had comments in site.pp [14:46:41] anyway, cool! [14:47:11] i'm drawing a complete blank, looking! [14:47:58] Jeff_Green: I now know enough to know that I don't care [14:48:00] no worries! [14:48:17] yeah but now I care [14:48:20] :-P [14:48:44] oh interestin [14:48:45] g [14:49:10] wow, somewhere in the recesses of my brain there's history on this [14:49:26] I fired up a mysql instance for some kind of testing [14:51:35] huh! [14:58:31] hashar: you've got a message in #wikimedia-labs ;) [15:00:38] Thehelpfulone: aka here :-D [15:00:51] I am not part of the ops team so I can't +2 operation/puppet changes [15:01:10] ok [15:01:10] you will have to find a op to take care of your changes :-] [15:01:13] but notpeter is? ;) [15:01:57] notpeter: ah, hume was used for testing mysql schema mods leading up to the civicrm upgrade, I believe I can tear down that stuff now that they're done [15:07:43] Thehelpfulone: yeah. someone gave this money access to the +2 button [15:08:08] they must have been under the influence when they gave it to you ;) [15:08:09] [15:42:22] https://gerrit.wikimedia.org/r/#/c/6727 and https://gerrit.wikimedia.org/r/#/c/6584/ if you can :) [15:10:28] if you could oblige notpeter ^ [15:14:59] Thehelpfulone: so, iwill definitely merge the second one, but I actually think that the first one could be done better. can you set it up so to just notify => Service["lighttpd"] instead ? [15:15:12] without the exec definition, I mean [15:15:17] !log updating firmware on storgae3 [15:15:20] Logged the message, RobH [15:15:40] (or if that's not reasonable, let me know why) [15:15:58] Jeff_Green: so i am going to update the drac firmware, then bios, then the raid controller [15:16:04] and see if i cannot clear the error [15:16:35] ok [15:17:01] notpeter: sure I'll have a go, I'm new to all this git gerrit stuff so just replace notify => Exec["service-lighttpd-reload"]; with notify => Service["lighttpd"] ? [15:18:39] yeah, although I guess that will do a restart instead of a reload [15:18:45] lemme look at the puppet docs a little [15:19:22] notify might not do a restart [15:19:29] i think it depends on how the service is set up [15:19:42] i think if it is hasreload => true [15:19:45] or someting like that [15:19:47] it would reload [15:19:50] that's just a guess though [15:22:56] ottomata: yeah, I'm trying to figure that out [15:22:59] oh puppet docs [15:23:05] you're so.... mildly ok [15:24:10] ottomata: ah, sadly, no.looks like there's a ticket to create such functionality in puppet [15:24:33] Thehelpfulone: okie dokie. I can merge that, as it doesn't seem like there's much of a better way. *sigh* [15:24:52] ok [15:25:50] aye rats [15:25:58] ok [15:26:01] there is another way though [15:26:04] if you really want to restart [15:26:08] instead of notifying the service [15:26:12] you can set up an exec that reloads [15:26:15] that gets notified [15:26:23] ottomata: yeah, that's what Thehelpfulone did [15:26:29] and it will totally work [15:26:36] but I thought there was a prettier way [15:26:37] I didn't do that, credits to jeremyb :) [15:26:42] ahh, ok [15:26:42] ah, ok [15:26:42] cool [15:27:06] actually, maybe better than notifying [15:27:09] is to subscribe the exec [15:27:29] well, hmm, that would only be better if you are only using this exec for one file [15:27:34] so you can do [15:27:43] subscribe => File[/whatever] [15:27:45] on the exec [15:27:51] with refreshonly => true [15:27:57] which will only run the exec if the file changes [15:28:12] OR, the conversly the file can notify the exec to reload, like i guess you are doing now [15:28:34] ja [15:28:38] should I re-run puppet now notpeter? [15:28:50] do either of you guys know about how to set up new nagios monitoring and/or triggers? [15:29:07] i'm not messing with it right now, but I might soon and I want to know how it is done [15:29:10]