[00:02:25] paravoid: 1.8M commons temp objects...graet [00:02:27] *great [00:02:36] this cleanup script will take ages [00:03:00] it would be faster to backport a change to make it do batching, sync, and restart [00:03:12] (03PS1) 10Dzahn: remove gurvin and gurvin.mgmt, decom [operations/dns] - 10https://gerrit.wikimedia.org/r/94448 [00:08:13] (03PS1) 10Dzahn: remove gurvin from dhcp, dsh and site.pp [operations/puppet] - 10https://gerrit.wikimedia.org/r/94449 [00:09:21] (03CR) 10Dzahn: "see comments on the linked ticket. just wondering why it still had certificates::wmf_ca on it." [operations/puppet] - 10https://gerrit.wikimedia.org/r/94449 (owner: 10Dzahn) [00:32:51] !log aaron synchronized php-1.23wmf2/maintenance/cleanupUploadStash.php '01da29d3ce23cea02df3fc975d6d51bfbc50a222' [00:33:11] Logged the message, Master [00:45:30] (03PS1) 10Dzahn: remove osm-web1-4 and mgmt [operations/dns] - 10https://gerrit.wikimedia.org/r/94457 [00:46:05] !log aaron synchronized php-1.23wmf3/maintenance [00:46:20] Logged the message, Master [01:11:17] !log aaron synchronized php-1.23wmf2/includes/filebackend '2afdc066f52b54faf63f9c980f7ef6a7841dd094' [01:11:36] Logged the message, Master [01:11:47] !log aaron synchronized php-1.23wmf3/includes/filebackend '2afdc066f52b54faf63f9c980f7ef6a7841dd094' [01:12:04] Logged the message, Master [01:16:51] !log gwicke synchronized php-1.23wmf2/extensions/Parsoid/php [01:17:05] Logged the message, Master [01:19:30] !log aaron synchronized php-1.23wmf3/maintenance 'c674d4c90c40d18187d893ed7a7ea94f43a1a624' [01:19:50] Logged the message, Master [01:19:53] and parsoid jobs are back: https://ganglia.wikimedia.org/latest/?r=hour&cs=&ce=&s=by+name&c=Parsoid%2520eqiad&tab=m&vn= [01:19:59] !log aaron synchronized php-1.23wmf2/maintenance 'c674d4c90c40d18187d893ed7a7ea94f43a1a624' [01:20:18] Logged the message, Master [01:23:15] !log gwicke synchronized php-1.23wmf3/extensions/Parsoid/php [01:23:30] Logged the message, Master [01:26:16] <^d> Um, what's going on with the apis? [01:26:17] <^d> http://ganglia.wikimedia.org/latest/graph_all_periods.php?c=API%20application%20servers%20eqiad&m=cpu_report&r=hour&s=by%20name&hc=4&mc=2&st=1383960343&g=network_report&z=large [01:26:32] <^d> (network, last hour) [01:27:01] ^d: parsoid back in action [01:27:07] <^d> Ah, gotcha :) [01:27:18] <^d> Doesn't look dangerous, just happened to notice. [01:27:21] <^d> :p [01:27:30] 99% of our load comes from our the job runner which I just fixed [01:27:43] <^d> *nod* [01:29:38] we do a lot of requests for content, but each of those is quite cheap for the api [02:03:45] !log LocalisationUpdate completed (1.23wmf2) at Sat Nov 9 02:03:45 UTC 2013 [02:04:05] Logged the message, Master [02:04:56] !log LocalisationUpdate completed (1.23wmf3) at Sat Nov 9 02:04:55 UTC 2013 [02:05:13] Logged the message, Master [02:10:54] !log LocalisationUpdate ResourceLoader cache refresh completed at Sat Nov 9 02:10:54 UTC 2013 [02:11:13] Logged the message, Master [02:28:53] (03PS1) 10Jforrester: Create visualeditor-default.dblist to simplify config [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94471 [02:28:53] (03PS1) 10Jforrester: Remove cruft from wmgVisualEditorDisableForAnons no longer needed [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94472 [02:32:51] (03PS5) 10Jforrester: Enable VisualEditor on board, collab, office and wikimaniateam wikis [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/93780 [03:00:24] ori-l: About [03:00:25] ? [03:01:00] Or someone else from ops.. [03:01:21] Could someone add me as a projectadmin to the tool labs project please? [03:42:37] PROBLEM - Puppet freshness on sq48 is CRITICAL: No successful Puppet run in the last 10 hours [04:57:49] (03PS2) 10TTO: Renamed $wmfConfigDir to $wmgConfigDir in mediawiki-config [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94354 (owner: 10Arav93) [04:57:53] (03CR) 10jenkins-bot: [V: 04-1] Renamed $wmfConfigDir to $wmgConfigDir in mediawiki-config [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94354 (owner: 10Arav93) [05:35:54] !log aaron synchronized php-1.23wmf2/includes/filebackend/SwiftFileBackend.php '519883bd67c960998e8cb2eb0cda52f3764192f7' [05:36:11] Logged the message, Master [05:37:47] !log aaron synchronized php-1.23wmf3/includes/filebackend/SwiftFileBackend.php '519883bd67c960998e8cb2eb0cda52f3764192f7' [05:38:05] Logged the message, Master [05:46:09] (03PS3) 10TTO: Clean up wgSiteName in InitialiseSettings [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/86418 [05:48:15] paravoid: https://blueprints.launchpad.net/swift/+spec/pool-memcache-connections sounds interesting [05:55:34] https://bugs.launchpad.net/swift/+bug/1229142 is fun too (not that it effects us) [06:10:16] Reedy: still around? [06:11:43] ah it looks like someone already added you [06:13:20] apergos: maybe you can merge https://gerrit.wikimedia.org/r/#/c/94480/1 :) [06:13:27] hahaha [06:13:44] * Aaron|home already backported it [06:14:07] ok but maybe you can look at https://gerrit.wikimedia.org/r/#/c/91213/ :-D [06:15:31] I see it in the db conf, ok [06:16:26] lgmt merged [06:16:35] *tm [06:20:12] oh man, more trademark bs? [06:20:40] which thing is this? [06:22:04] (03CR) 10ArielGlenn: [C: 031] fix pdf servers in dsh groups [operations/puppet] - 10https://gerrit.wikimedia.org/r/94307 (owner: 10Dzahn) [06:22:30] apergos: you :) [06:22:35] apergos: the "*tm" [06:22:41] sorry, bad late-ish night joke [06:22:43] :-P [06:22:48] it is, and it's early morning [06:23:01] I"m awake enough to review (a little) but not awake enough for that! [06:24:12] nt awake enough to review bugz module though... later or monday [07:08:16] RobH: 2 ideas from the people sitting next to me: archivists and whistleblowers [07:22:28] then we will get to have he debate about whether a misc host named 'snowden' is too obvious a target for script kiddies [07:22:57] archivists might overlap some with encyclopedians [07:34:58] keep the private wiki dumps on snowden >.> [07:40:19] FAT CHANCE [07:40:47] no fun [07:41:04] dang straight [08:01:04] PROBLEM - Puppet freshness on sq80 is CRITICAL: No successful Puppet run in the last 10 hours [08:02:52] snowden assange and manning could be the search boxes... [08:47:29] (03CR) 10Nemo bis: "(See steps at .)" [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94354 (owner: 10Arav93) [08:51:53] morning [08:52:48] mogge [08:53:06] springle: hey [08:53:24] can I bother you for 3' to give you an update on something? :) [08:55:10] paravoid: sure (apergos started already) [08:56:04] hello [08:56:20] we had an outage yesterday [08:56:26] between 18:17 UTC and 18:31 UTC [08:56:56] s1 shard had a huge load spike, along with number of rows, open transactions, rows select [08:57:02] cause is still unknown [08:57:16] we lost db1050 in the process, it just became completely unresponsive [08:57:44] I verified that load started increasing across the databases before we lost db1050, though, so it's likely an effect, not a cause [08:57:59] (although it then obviously made the situation even worse) [08:58:19] we powercycled db1050 and then something wrong with LVM and hasn't come up properly yet -- I depooled it from mw10NN [08:59:02] I digged through ishmael, found a few outliers and pointed them out, one of them was fixed but it doesn't look related to this [08:59:17] ori also found an uploadwizard slow query and turned it off [08:59:38] no hard evidence it was the cause of this [09:00:36] interesting that unpurged txns on db1056 master spiked over that time. something wrote heavily to it (which binlog may show) or else ran slow SELECT causing purge lag [09:01:31] oh and https://ishmael.wikimedia.org/more.php?hours=24&host=db1052&checksum=2279768281867546317 was the other one that I found that looked nasty [09:01:47] at one point, it was 9% of queries, 90% of time for past 1 hour [09:02:05] and it does look elevated since yesterday [09:02:33] and during a very similar time to the outage (20' before) [09:02:35] yes LogPager has been a problem for a while. bug open, and believe it or not it's actually better than it was [09:02:51] that's what I said ;) [09:03:14] Aaron|home: indeed, sorry, I didn't mean to disregard your comment, just dumping all the information to sean :) [09:03:47] paravoid: thank you :) [09:04:28] I was thinking of writing the incident report, but it's all so unclear that it'd look ridiculous [09:04:44] paravoid: heh, I was just talking about the "actually better than it was" bit [09:04:55] "something happened somewhere and we crashed" [09:05:09] were the other layers flapping because of db failures, or in advance of it? [09:05:37] my investigation showed a cascading failure starting from databases [09:05:45] ok [09:06:02] i.e. apache requests piled up waiting for dbs and produced a site-wide (all shards) failure [09:06:09] both regular appservers and api appservers [09:11:03] there was definitely a write spike on the S1 master db1056 at that time [09:12:27] how do you figure? [09:13:08] can you point me at what you're looking? [09:13:18] so that I know for next time ;) [09:15:12] Aaron|home: btw, I got a tentative window from greg-g next thursday, to switch the Swift master [09:16:13] paravoid: http://ganglia.wikimedia.org/latest/graph.php?r=day&z=xlarge&title=mysql_innodb_transactions_unpurged&mreg[]=^mysql_innodb_transactions_unpurged%24&hreg[]=db1056&aggregate=1&hl=db1056.eqiad.wmnet|MySQL+eqiad [09:16:36] two spikes, first one about the right time? [09:16:48] yes [09:17:04] when there's doubt, I use the CSV button on top of the graph [09:17:13] then go the timestamp and manually read it [09:17:39] (what I'd give for interactive graphs you can hover over and they'd tell you the timestamp & value) [09:19:32] similar metric, but may also indicate slow SELECT on master: http://ganglia.wikimedia.org/latest/graph.php?r=day&z=xlarge&title=mysql_innodb_history_list&mreg[]=^mysql_innodb_history_list%24&hreg[]=db1056&aggregate=1&hl=db1056.eqiad.wmnet|MySQL+eqiad [09:21:26] looks like ganglia didn't get data from many S1 boxes around that time, but innodb hist list length takes time to reduce after a spike, so we can see the aftermath [09:33:32] PROBLEM - MySQL InnoDB on db1002 is CRITICAL: CHECK_NRPE: Socket timeout after 10 seconds. [09:34:21] RECOVERY - MySQL InnoDB on db1002 is OK: OK longest blocking idle transaction sleeps for 0 seconds [09:59:54] springle: are you still investigating? I want to play with ganglia a bit and I may make the web interface unavailable in the process [10:00:00] but I can wait if you are [10:00:16] paravoid: i am, but that's fine. (am in binlogs) [10:08:42] !log upgraded ganglia-web to 3.5.10 [10:09:02] Logged the message, Master [10:31:14] paravoid: link to new shiny features? :) [10:32:48] hmm several new small useful things I see [10:37:20] yeah, and breaking other crap [10:39:57] (03PS3) 10Ori.livneh: Add Graphite module & role [operations/puppet] - 10https://gerrit.wikimedia.org/r/92271 [10:40:22] grrr [10:40:31] the ganglia-web code is so horrible [10:43:52] too tired to write a proper commit message [10:44:06] before i go to bed, here's one i thought you might enjoy, paravoid [10:44:16] the actual, no-joke output of 'uwsgi --help': https://dpaste.de/N8js/raw/ [10:44:40] hahahaha [10:45:05] --spooler-harakiri *** UNDOCUMENTED OPTION *** [10:45:06] --mule-harakiri *** UNDOCUMENTED OPTION *** [10:45:10] rofl [10:45:34] yeah, i am actually laughing too, i mean -- seriously [10:45:55] --zerg-server *** UNDOCUMENTED OPTION *** [10:46:20] anyways, good night [10:46:36] paravoid: looking throught dberror.log during the outage; odd that connection issues are logged for many different mysql boxes, including external storage hosts(!?), by no means just S1 or wikidata, around the time [10:47:36] bye ori-l, thanks for the laugh [10:47:40] how would a purely S1 spike or outage would affect all the others? [10:47:45] springle: at what times where those though? [10:48:08] between 18:15 and 18:29 [10:48:12] we do know that apaches went haywire at some point, possibly due to an effect of the s1 overload [10:48:14] same period [10:48:17] requests piled etc. [10:48:41] so if it's the middle of the outage, it could be explained this way [10:48:50] earliest db errors logged in that period were not S1 hosts... [10:48:53] hmm ok [10:48:54] if it's as early as 18:15, though, then it won't explain it, no [10:53:29] paravoid: did anything happen to ganglia in that period? [10:53:44] or to stats aggregation in general [10:53:58] not afaik [10:57:37] springle: i'm thinking about using externallinks in labs. just wondering if that's one that will be getting a primary key or modified timestamp. (you were doing that for other tables, right?) [10:57:51] or maybe you have extra columns already and i just can't see them [10:59:03] jeremyb: externallinks is getting a PK now. it's about half done. a few of the largest wikis remain to migrate [10:59:30] springle: woot. any particular way i should track that? !log ? [11:00:30] jeremyb: i !log-ged the start of the process, but not every wiki [11:00:51] ok, maybe i'll just look or ask again in a bit [11:01:28] i guess the new PK is guaranteed higher (it's an insert + autoincrement) if it's removed and readded [11:01:42] so that tells you chronology but not actual time [11:01:54] it's going to be days, likely a week for enwiki, to avoid lagging slaves. and I'm only running it while I can be around to watch dbtree ;) [11:02:07] right [11:02:28] i suppose i could start with a smaller one while i get it working. enwiki was the original target [11:03:09] once we have PK we can do online schema migrations much faste. no technical problem with having an el_stamp field. is there a bug or gerrit commit for it already? [11:03:47] no. haven't even discussed it with anyone. i thought maybe you were giving it to me for free ;) [11:03:55] heh [11:03:59] no such luck [11:04:28] a related use case I might address with this: https://bugzilla.wikimedia.org/32026 [11:04:35] but not the original reason i was thinking of it [11:26:14] springle: so is there 1 or 2 DBs already finished I can test against? (externallinks PK) [11:30:31] jeremyb-phone: many. bgwiki for eg. but not in labs yet i suspect, as it would require rebuilding views [11:31:00] this I can test ;) [11:39:22] springle: you're right, it's in bgwiki but not bgwiki_p (not the view) [11:44:00] paravoid: https://ishmael.wikimedia.org/more.php?host=db1052&hours=24&checksum=10298050210183520632 [11:46:07] was that one you guys already discussed? [11:50:15] springle: yes [11:50:32] I pointed it out, noone could tell me what it does, I left it at that [11:52:57] i think you were right to wonder about it [11:55:34] huh.. the sample ishmael captured shows LIMIT 86991,2 [11:56:02] another one LIMIT 86779,2 [11:56:05] etc [12:08:03] it's increased in frequency in the last week or so too. several short bursts of many concurrent instances each day [13:07:19] (03CR) 10Helder.wiki: [C: 031] Make VisualEditor namespaces extend, not replace, default [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94420 (owner: 10Jforrester) [13:42:49] PROBLEM - Puppet freshness on sq48 is CRITICAL: No successful Puppet run in the last 10 hours [13:46:30] !log starting pt-kill daemon on S1 slaves for SpecialAllpages::showToplevel queries > 30s [13:46:46] Logged the message, Master [14:30:42] anyone around to help with RT/download.wm.o [14:31:11] ? [14:31:16] https://rt.wikimedia.org/SelfService/Display.html?id=6270 [14:33:06] (03PS1) 10Arav93: Renamed $wmf* to $wmg* to improve consistency [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94503 [14:33:08] (03CR) 10jenkins-bot: [V: 04-1] Renamed $wmf* to $wmg* to improve consistency [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94503 (owner: 10Arav93) [14:34:08] (03CR) 10Anomie: [C: 04-1] "(1 comment)" [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94354 (owner: 10Arav93) [14:38:50] (03CR) 10Anomie: "Also, this appears to be roughly identical to I5a18a70a. One or the other should be abandoned," [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94354 (owner: 10Arav93) [14:38:53] (03CR) 10Anomie: [C: 04-1] "(1 comment)" [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94503 (owner: 10Arav93) [15:33:19] (03Abandoned) 10Arav93: Renamed $wmf* to $wmg* to improve consistency [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94503 (owner: 10Arav93) [15:38:06] (03PS3) 10Arav93: Renamed $wmfConfigDir to $wmgConfigDir in mediawiki-config [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94354 [15:38:13] (03CR) 10jenkins-bot: [V: 04-1] Renamed $wmfConfigDir to $wmgConfigDir in mediawiki-config [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94354 (owner: 10Arav93) [16:38:58] Thehelpfulone: didn't we have a bug report about timeouts on mailman when dealing with spam/moderation queue? [16:39:22] I remember eia and/or Fae complaining about it [16:39:43] oh well, Nikerabbit will file another one [16:40:35] Nikerabbit: I think it's unrelated [16:40:39] Dario says "FYI there’s a temporary issue that prevents people from unsubscribing from this list. I notified ops." [16:40:45] that's probably your problem [16:41:30] Nemo_bis: where when what list [17:08:02] (03PS1) 10Nemo bis: Make current $wgNoFollowLinks = true explicit [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94508 [17:42:07] Hi, I'd like to parse through server debug logs for "edit conflict" stuffs. Is there a policy around how to pursue this? And, for how long are debug logs archived? [17:43:06] awight: You'll need to start collecting them first [17:44:06] Reedy: meaning just that they are split by appserver? [17:53:06] awight: do you know https://wikitech.wikimedia.org/wiki/Logs ? [17:53:23] (03CR) 10Aklapper: [C: 04-1] "Patch is completely broken with lots of ">>>>>>>" noise (probably due to a missing rebase)." [operations/mediawiki-config] - 10https://gerrit.wikimedia.org/r/94354 (owner: 10Arav93) [17:53:36] better link for hexmode https://rt.wikimedia.org/Ticket/Display.html?id=6270 :) [17:54:02] i was momentarily confused and thought that RT had somehow been "upgraded" :P [17:54:16] Nemo_bis: interesting, thanks for that link. I'll poke around to see if wfDebug goes into one of these buckets. [17:54:46] jeremyb: tyvm [17:55:51] awight: is that log supported in MediaWiki? if yes you just need to check the config to see if it's enabled; and if yes, AFAICS every log group has its own file [17:57:55] Nemo_bis: i think he was asking if they are split into different files (or stored on different machines) per apache box [17:58:16] that's the old question, which the page should have answered [17:59:11] Nemo_bis: sure, just distinguishing from "log group has its own file" [18:01:41] PROBLEM - Puppet freshness on sq80 is CRITICAL: No successful Puppet run in the last 10 hours [18:04:03] Nemo_bis: ah, so it seems that wfDebugLog() dumps into those buckets, but I still don't see what happens to plain old wfDebug() messages [18:07:54] awight: they're not logged [18:08:17] boo. K, thanks for the confirmation [18:08:38] awight: do you have access to flourine? [18:08:52] that's my understanding, it's all hearsay for me :) [18:10:40] jeremyb: yep [18:13:21] currently causing a processor to die of grep ;) [18:14:22] die may be inaccurate :) [18:14:32] but i got it [18:14:57] or the disk, rather [18:16:04] rats. Looks like Nemo_bis is right that wfDebug > /dev/null [18:17:26] awight: what's the size of all those logs currently, btw? [18:17:58] mmm [18:18:07] they are rotated daily, i believe [18:18:33] seems that today's logs add up to 41GB [18:18:54] and the archive is 506GB (so perhaps rotation is less frequent?) [18:19:08] get a usb key ready ;) [18:20:05] thanks [18:21:04] Nemo_bis, I thought it was just an intermittent issue with Chrome? [18:21:20] never heard of this [18:21:30] though yes, Chrome is horrible ;) and yes, he uses it [18:23:00] yeah generally Mailman and Chrome have issues from time to time [18:23:13] I've had problems with time outs etc too, even when trying to modify list settings [18:24:00] I think Mailman said words to the effect of "don't use Chrome" [18:25:44] I see [18:26:19] what's wrong with the SMTP interface? [18:26:39] blank mail to ${listname}-unsubscribe@lists.wikimedia.org [18:34:07] wondering what this means https://ganglia.wikimedia.org/latest/graph.php?h=fluorine.eqiad.wmnet&m=cpu_report&r=year&s=by%20name&hc=4&mc=2&st=1384022007&g=network_report&z=medium&c=Miscellaneous%20eqiad [20:57:00] PROBLEM - MySQL Slave Running on db1047 is CRITICAL: CHECK_NRPE: Socket timeout after 10 seconds. [20:57:10] PROBLEM - MySQL Recent Restart on db1047 is CRITICAL: CHECK_NRPE: Socket timeout after 10 seconds. [20:57:50] RECOVERY - MySQL Slave Running on db1047 is OK: OK replication Slave_IO_Running: Yes Slave_SQL_Running: Yes Last_Error: [20:58:00] RECOVERY - MySQL Recent Restart on db1047 is OK: OK 2698463 seconds since restart [22:35:59] MaxSem: did the UW disabling expensive queries patch get merged/ [22:36:09] ? [22:36:45] nope, because we didn't want to change the situation when there was higher-priority stuff to investigate [22:37:18] but it should - in graphite it ranges somewhere between 10 and 100 seconds [22:37:33] MaxSem: *seconds*!? [22:37:34] wat [22:38:08] https://graphite.wikimedia.org/dashboard/temporary-35 [22:39:16] MaxSem: i see some really tiny dots. what are the units for the Y axis? [22:39:35] ms [22:39:57] ugh, ok [22:47:02] home sweet home, YuviPanda? :) [22:47:13] MaxSem: don't get me started :P [22:57:25] hey YuviPanda [22:57:34] hi ori-l :D [22:57:34] glad you made it home safe [22:57:53] ori-l: yeah, and thanks to a long sandwich before the flight didn't have to eat shitty flight food :D [22:58:58] heh [23:06:03] MaxSem: feel free to disable UW's expensive stuff if that's causing trouble, I can find a solution for it later on [23:06:35] will do on Tue as part of deployment training [23:06:45] MaxSem: ah, okay [23:06:49] MaxSem: can you merge https://gerrit.wikimedia.org/r/#/c/94586/? [23:06:53] trivial trivial trivial [23:06:57] i would merge it myself but COI [23:07:00] :P [23:07:56] done [23:08:03] ty MaxSem [23:08:07] MaxSem: whom are you training? [23:08:35] MaxSem: maybe I should join in on the training too, even though I don't have deploy rights [23:08:55] it's for those who don't have em [23:09:17] MaxSem: can I recommend http://asciinema.org/ ? [23:09:26] you can record your terminal session really easily using that tool [23:09:32] MaxSem: aha! [23:09:45] you can install it locally and then ssh [23:09:57] we will do it via hangouts [23:10:27] are you going to record it? [23:10:40] the reason I'm asking is twofold -- one, there are folks like bd808 that need some training too [23:10:54] ohai bd808 [23:11:02] two, it'd be useful to hear you describe what works and what sucks about the process [23:11:14] since core platform & ops are working on improving it [23:11:36] if it's gonna be git-deploy, what's the matter? [23:12:14] there's a lot of the deployment process that git-deploy will simply invoke via hooks [23:12:20] * bd808 waves his untrained n00b hand at MaxSem  [23:12:27] but how that stuff works and what it does and what logs it emits, etc hasn't been determined yet [23:13:24] bd808, wanna join us? [23:14:06] MaxSem: Sure. [23:15:45] MaxSem: What time on Tuesday? [23:16:31] 2-4pm, see your inbox [23:17:11] now for realz [23:17:12] MaxSem: yay thanks! [23:17:25] * bd808 waits for gmail which is sloooow [23:17:50] it's me who's a gcal noob [23:21:56] l33t gcal skillx0rs [23:22:55] MaxSem: Got it. Overlaps with Core weekly but watching deploy stuff sounds more entertaining [23:43:39] PROBLEM - Puppet freshness on sq48 is CRITICAL: No successful Puppet run in the last 10 hours