[01:30:11] <wm-bot>	 !log tools.slf <legoktm> set `git config user.{name,email}` and then manually triggered crontab
[01:30:13] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.slf/SAL
[06:03:15] <legoktm>	 revi: yes, I assume any domain change policy has backwards-compatibility as a part of it. "Cool URIs don't change"
[16:06:44] <bd808>	 revi: yes. Not explicitly written in that proposal, but we would implement some compatibility system for the tools.wmflabs.org and *.wmflabs.org URLs that are active at the time we start switching things. We still maintain toolserver.org URL forwarding for similar reasons.
[16:08:36] <revi>	 thanks
[17:38:34] <bstorm_>	 !log tools Adding the prometheus servers to the ferm rules via wikitech hiera for kubelet stats T228573
[17:38:38] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL
[17:38:38] <stashbot>	 T228573: toolforge k8s nodes oom? - https://phabricator.wikimedia.org/T228573
[17:40:17] <lucaswerkmeister>	 has anyone managed to run a Jetty server on Toolforge? since the Stretch update (I think), I keep running into the process limit
[17:40:46] <lucaswerkmeister>	 (`ulimit -u` = 30) – either Java crashes in different ways, or it eats up all the available processes so that nothing else can run alongside it
[17:55:39] <bstorm_>	 !log tools draining tools-worker-1023 since it is having issues
[17:55:41] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL
[17:57:57] <lucaswerkmeister>	 wtf, Java somehow behaves differently if I precede it with `ulimit -u 30`, even though 30 is already the current value
[18:04:34] <lucaswerkmeister>	 I don’t understand how that’s possible
[18:09:16] <lucaswerkmeister>	 (hm it looks like it’s actually nondeterministic and just happened to line up a few times)
[18:14:15] <bd808>	 lucaswerkmeister: java has been a pain on the Stretch job grid if that's where you are trying to get jetty working
[18:14:35] <lucaswerkmeister>	 right now it’s login-dev but ultimately it would be the grid yeah
[18:15:14] <lucaswerkmeister>	 but I’m looking into something else now anyways – ultimately I need a simple SPARQL server for a plain n-triples file on disk
[18:15:26] <lucaswerkmeister>	 and if Apache Jena Fuseki is being a pain, I’ll try to use Python rdflib instead :)
[18:15:32] <lucaswerkmeister>	 let’s see where that gets me…
[18:15:38] <bd808>	 ok. that makes the ulimit stuff make sense too. We have the bastions tightened down to try and keep folks from running real tools there
[18:15:59] <lucaswerkmeister>	 does the grid have different limits? because I think I saw a similar error there as well
[18:17:02] <lucaswerkmeister>	 oh, no, that was a completely different error
[18:17:10] <bd808>	 yeah, the job grid is ... different. :) Grid Engine doesn't understand cgroups (or at least our deploy doesn't) so limits are very different there
[18:17:11] <lucaswerkmeister>	 then let me try if I’ve just been wasting my time ^^
[18:17:52] * bd808 tries to grok "simple" and SPARQL in the same sentence
[18:18:34] <lucaswerkmeister>	 no UPDATE, don’t care much about performance (simple queries expected)
[18:28:12] <bstorm_>	 !log tools depooled tools-sgewebgrid-lighttpd-0902 to find out why it is behaving weird
[18:28:15] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL
[18:32:51] <bstorm_>	 !log tools repooled tools-sgewebgrid-lighttpd-0902 after restarting the grid-exec service
[18:32:52] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL
[18:33:17] <bstorm_>	 !log tools depooled tools-sgeexec-0905 because it's acting kind of weird and not responding to prometheus
[18:33:19] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL
[18:39:53] <bstorm_>	 !log tools repooled tools-sgeexec-0905 after reboot
[18:39:56] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL