[01:31:29] <legoktm>	 James_F: a patch appears
[01:32:26] <legoktm>	 I'd still want M.oritz or someone more experienced with reprepro to review it though
[01:32:53] <James_F>	 legoktm: Thanks! :-)
[02:07:38] <sukhe>	 legoktm: wow, thanks! (I am no mo.ritz but I can have a look tomorrow morning as I have added stuff to reprepro :)
[02:08:04] <legoktm>	 awesome
[09:17:42] <elukey>	 hello hello
[09:18:37] <elukey>	 it happens almost daily that some ores nodes get confused after a sighup, triggered by logrotate
[09:19:02] <elukey>	 all uwsgi workers go down, but then they don't restart
[09:19:17] <elukey>	 puppet usually fix this when the next puppet run comes
[09:19:35] <elukey>	 logrotate uses copytruncate + uwsgi reload, is necessary?
[09:19:56] <_joe_>	 what are you asking?
[09:20:03] <_joe_>	 if the sighup is necessary?
[09:20:18] <elukey>	 yeah
[09:21:10] <_joe_>	 the sighup comes from uwsgi reload, correct?
[09:21:11] <elukey>	 in /etc/logrotate.d/uwsgi we use only copytruncate, but in /etc/logrotate.d/ores there is also a reload
[09:21:19] <_joe_>	 oh I see
[09:21:27] <_joe_>	 ok, changing that seems sensible
[09:22:00] <elukey>	 the logrotate is the shared one for service::uwsgi I think, so either we add a specific flag to test only on ores or we change it for all
[09:22:19] <elukey>	 that should be fine in theory, we don't have that many uwsgi things around
[09:27:01] <elukey>	 (sending a cr)
[16:17:43] <effie>	 jbond42: should I merge your change?
[16:17:56] <effie>	 John Bond: add cert (b0e185c)
[16:18:42] <jbond42>	 ahh yes please is just priv repo
[16:22:30] <volans>	 "just"™
[21:10:37] <mutante>	 legoktm wins the "most signatures on a GPG key" award with "129 signatures unchecked due to missing keys" in addition to the other SRE
[21:13:29] <legoktm>	 hehehe :D
[21:14:21] <legoktm>	 most are from keysigning parties at linux.conf.au, all hands and DebConf
[21:17:37] <mutante>	 I am currently trying to add you to pwstore and the usual dance of some other keys being invalid
[21:49:04] <apergos>	 isn't it easier now that the keys are local? or is that not helping?
[21:52:48] <mutante>	 no, it's not :)
[21:52:56] <apergos>	 oh dear :-D
[21:53:12] <mutante>	 at least ..not yet
[21:53:36] <mutante>	 unless we build the .keyring file from the keys in ./keys/ and keep them updated
[21:53:52] <mutante>	 and don't realy on "pws update-keyring" updating it from keyservers
[21:54:56] <sukhe>	 the problem with that being that it results in missing keys because it points to keys.openpgp and that's not federated?
[21:55:45] <mutante>	 every time you run it with a different keyserver, you get a different set of keys that it does not find
[21:55:55] <mutante>	 but if you repeat it enough times with enough keyservers
[21:56:02] <mutante>	 eventually you get the complete collection
[21:56:17] <mutante>	 then you must be lucky none are expired or revoked
[21:56:17] <sukhe>	 ah :|
[22:03:24] <mutante>	 I pushed. It should work for legoktm now.
[22:13:02] <rzl>	 well, that's a new-to-me popup from gerrit, after clicking submit https://phabricator.wikimedia.org/P13598
[22:13:11] <rzl>	 seems to have merged anyway
[22:16:34] <sukhe>	 sorry rzl for the possible nitpicking :( I didn't realize there was a +1 from dzahn until I submitted it
[22:16:50] <rzl>	 sukhe: no you're good, thanks for calling it out -- I was typing too fast :P
[22:17:12] <rzl>	 and thanks for the quick reviews, too!
[22:17:16] <sukhe>	 this to me is the definition of nitpicking and I should be ashamed :P
[22:22:24] <rzl>	 I either put the typo back in by accident because I pushed from my local repo after fixing it in gerrit... or I put it back in on purpose as a personal test for sukhe, believe what you will
[22:22:36] <sukhe>	 rzl: you are too kind
[22:22:56] <rzl>	 (but genuinely, I'd rather have it correct, so thanks for pointing it out, please keep doing it)
[22:23:57] <legoktm>	 woot! let me give pwstore a try now
[22:24:51] <mutante>	 legoktm: yes, please git pull and then   "pws ed" a random file
[22:24:59] <mutante>	 well, a random one for the "ops" group
[22:25:38] <cdanis>	 `pws ed management` is a good test :)
[22:26:38] <mutante>	 or maybe one that is easier to change
[22:26:49] <mutante>	 like mailman-passwds
[22:27:39] <legoktm>	 it works, I'm able to read them
[22:27:56] <mutante>	 ok, great! then you can already check the box, basically
[22:28:04] <mutante>	 it's always "ed" for both reading and writing
[22:29:06] <apergos>	 which is maybe not the best but changing that is lowest priority
[22:30:14] <mutante>	 legoktm: take a look at the .users file. It maps the user to keys and at the bottom there are groups with users. so you are in "ops" and there is also "netops", "frackops" and "dcops". Most files are encrypted for ops and a few special cases for the other groups.
[22:30:45] <mutante>	 If you feel the need for one of those then please ask one of the existing members.
[22:31:09] <legoktm>	 ok, and then if I wanted to change or add something, I `pws ed ...` it, save, and then git commit and push the result?
[22:32:23] <mutante>	 legoktm: yes
[22:33:14] <mutante>	 https://office.wikimedia.org/wiki/Pwstore#Basic_usage_commands
[22:36:03] <legoktm>	 ack