[10:05:48] <dcausse>	 lunch
[13:12:22] <inflatador>	 <o/
[13:15:58] <dcausse>	 o/
[13:20:15] <inflatador>	 CODFW is now 100% OpenSearch!
[13:21:41] <dcausse>	 nice!
[13:21:53] <inflatador>	 I just merged your patch to re-enable the indices
[13:23:10] <dcausse>	 thanks!
[13:42:33] * inflatador also needs to change the concurrent shard recoveries back down to 4
[14:00:34] <ebernhardson>	 \o
[14:03:04] <inflatador>	 \moti wave
[14:03:07] <inflatador>	 <o/
[14:03:39] <inflatador>	 dcausse sorry I'm late, just joined pairing if you are interested
[14:03:50] <dcausse>	 sure
[14:03:55] <dcausse>	 o/
[14:05:59] <dcausse>	 ebernhardson: we're discussing envoy if you wanna join https://meet.google.com/aod-fbxz-joy?authuser=0
[14:06:07] <ebernhardson>	 sure
[14:06:40] * ebernhardson was pondering if i care that character class ranges can include \uE000 and \uE0001...
[14:48:22] <ebernhardson>	 i did find a response in 2019 from someone who seems knowledgable, saying that 50 node master elections are still undesirable. 
[14:49:06] <ebernhardson>	 Essentially the election works on timeouts, where it expects to get votes in a certain timeframe. If that fails it runs another election with a longer timeframe, repeatedly.  Doing that on a large number of nodes might timeout multiple times slowing down elections
[14:49:26] <ebernhardson>	 https://discuss.elastic.co/t/whats-the-downside-of-making-all-data-nodes-master-eligible/175495/2
[14:52:49] <inflatador>	 👀
[15:57:55] <inflatador>	 working out, back in ~40
[16:42:37] <inflatador>	 back
[17:50:43] <inflatador>	 ebernhardson do we need to restart the cirrussearch nodes after we upgrade the madvise pkg?
[18:02:26] <inflatador>	 lunch, back in ~40
[18:06:16] <ebernhardson>	 inflatador: nope, it works external to elastic
[18:06:23] <ebernhardson>	 s/elastic/cirrussearch nodes/
[18:18:55] <ebernhardson>	 hmm, not finding a shared utility library between search-extra and cirrus-highlighter. And gerrit is down so can't check there :P
[19:40:46] <inflatador>	 sorry, been back awhile
[20:05:40] <dcausse>	 ebernhardson: should be https://gerrit.wikimedia.org/r/q/project:wmf-jvm-utils
[20:06:00] <ebernhardson>	 dcausse: thanks!
[20:15:36] <inflatador>	 OK, opensearch-madvise 0.2 is deployed on cirrus/cloudelastic/relforge. Checking to see if the systemd pieces are working...
[20:21:52] <inflatador>	 Looks like the unit file needs a tiny update, will get a patch up for that
[20:54:03] <inflatador>	 hmm, the command looks right in Puppet ( https://gerrit.wikimedia.org/r/plugins/gitiles/operations/puppet/+/refs/heads/production/modules/profile/manifests/opensearch/cirrus/server.pp#124 ), but the `ExecStart` is rendering without the 2nd arg (base_data_dir). Maybe the var isn't resolving correctly. Checking...
[21:04:25] <inflatador>	 ryankemper do you have anything for pairing? I'm just checking out the ExecSearch stuff
[21:04:58] <ryankemper>	 inflatador: nah I'm just looking over some graph split stuff and then gonna send out a couple comms
[21:09:40] <inflatador>	 ryankemper ACK, do you want to cancel then?
[21:09:53] <ryankemper>	 inflatador: yeah
[21:13:11] <inflatador>	 ryankemper gotcha. I'll be around for another hr or so, ping me if you need anything
[22:03:28] <inflatador>	 OK, I'm out. ryankemper if you wanna kick around the madvise stuff, I posted some updates in https://phabricator.wikimedia.org/T390592#10782249 . Otherwise, I'll take a look tomorrow
[22:03:47] <ryankemper>	 inflatador: ack
[22:05:14] <inflatador>	 cool, I saw a bunch of WDQS alerts that resolved...I assume you were doing a deploy? If not, may wanna keep an eye on that