[07:51:03] <ejoseph>	 good morning all
[07:51:26] <ejoseph>	 I am back, thanks for checking in on me
[07:56:17] <dcausse>	 ejoseph: hey, glad to see you feel better!
[08:00:01] <ejoseph>	 thanks, David
[10:00:23] <dcausse>	 lunch
[10:05:19] <gehel>	 Lunch
[10:09:06] <ejoseph>	 Hello everyone, I have a copy of the feedback I am submitting to the ERC 
[10:09:09] <ejoseph>	 https://docs.google.com/document/d/1C5V0b8DKqx8dlQ9u3A5bsfYLmQ1sBeZqIUEx4DTnpSo/edit?usp=sharing
[10:09:53] <ejoseph>	 Kindly review to see if I am misrepresenting  or omitted anything
[12:34:14] <gehel>	 ejoseph (and all): I've added a few more comments to that doc
[12:35:00] <gehel>	 @all: we're already past the deadline to review this doc. Could you make sure you review it and let Emmanuel and myself know when it's done? Before today - end of day?
[12:41:29] <dcausse>	 ejoseph: all good from me, thanks for working on this!
[13:07:47] <inflatador>	 greetings
[13:33:42] <dcausse>	 o/
[13:34:20] <dcausse>	 inflatador: do you know much time it would take to upgrade codfw to elastic7? 
[13:35:21] <dcausse>	 I haven't been following recent cluster rolling restarts so can't judge how long it will last
[14:05:27] <inflatador>	 dcausse Let me check my notes, but with the current tooling I think it's something like 5-6 hours
[14:05:49] <dcausse>	 oh wow that's fast
[14:11:04] <dcausse>	 inflatador: we want to reimage to bullseye all nodes before upgrading to es7 right?
[14:13:55] <inflatador>	 dcausse yes, we should be done with the bullseye upgrades within the next few wks
[14:14:07] <dcausse>	 cool, thanks!
[14:19:37] <dcausse>	 can we close T301959 ?
[14:19:38] <stashbot>	 T301959: Upgrade Search elasticsearch cluster / eqiad to elasticsearch 6.8.23 - https://phabricator.wikimedia.org/T301959
[14:22:33] <dcausse>	 we should probably close T218995 too I think
[14:22:33] <stashbot>	 T218995: re-enable deprecation warning logger on elasticsearch once issues are solved - https://phabricator.wikimedia.org/T218995
[14:23:36] <dcausse>	 closing those, feel free to reopen them if I'm wrong
[14:23:40] <inflatador>	 For T301959 , have we addressed the question at the bottom of the ticket?
[14:23:51] <inflatador>	 Other than that I'm totally fine with whatever
[14:24:15] <dcausse>	 looking
[14:26:40] <dcausse>	 ok done
[14:28:03] <dcausse>	 filed T308676, I tried to list all known prerequisites and draft a plan, please fix/amend anything you think is wrong or missing
[14:28:04] <stashbot>	 T308676: Elasticsearch 7.10.2 rollout plan - https://phabricator.wikimedia.org/T308676
[14:29:01] <dcausse>	 we could pack Week 2 and Week 3 if we're really fast rolling out elastic on codfw but not sure we should pressure ourselves here
[14:29:20] <dcausse>	 I mean pack Week 1 with Week 2
[14:29:53] <dcausse>	 errand, back later today
[14:36:43] <inflatador>	 OK, will take a look
[15:01:50] <gehel>	 ryankemper, mpham, Trey314159 : Wednesday meeting, and Tyler is here to talk with us!
[15:20:52] <mpham>	 i'm in a product meeting at the same time
[15:43:41] <inflatador>	 ebernhardson was that from the Stretch decoms last wk? Let's follow up offline if so
[15:57:58] <inflatador>	 workout, back in ~30
[16:39:00] <inflatador>	 back
[17:02:07] <gehel>	 ryankemper, inflatador, ebernhardson, Trey314159: can you have a look at the ERC feedback doc and confirm you are ok with it ? See email for details.
[17:04:27] <inflatador>	 gehel thought I did that already? Will look again
[17:08:22] <inflatador>	 I added a small note, but confirming again that I have read/understood
[17:38:44] <inflatador>	 lunch, back in 30-45
[18:13:53] <inflatador>	 back
[18:28:27] <gehel>	 inflatador: there is another document that Emmanuel shared today, with notes from our discussion. Those notes will be shared with the ERC Core team
[18:29:12] <inflatador>	 gehel thanks, I did miss that one, reviewing now...
[19:15:43] <inflatador>	 quick errand, back in 30
[19:35:37] <gehel>	 dcausse, ebernhardson: a question you might be able to answer: https://lists.wikimedia.org/hyperkitty/list/wiki-research-l@lists.wikimedia.org/thread/VXTSXKSG6WGEPHWJRTQQOHFV53H3LPUD/
[19:47:15] <ryankemper>	 \o
[19:48:25] <ryankemper>	 inflatador: there's a couple tickets, https://phabricator.wikimedia.org/T308339 and https://phabricator.wikimedia.org/T308331, that will require us to ban/depool a couple elastic hosts so that they can be reracked. It sounds like it won't happen for several months (https://phabricator.wikimedia.org/T308339#7937642) so there's no action required for us now, but I went ahead and added you as a subscriber to those two tasks as well
[19:55:51] <ebernhardson>	 gehel: i dunno, noone ever asked for the models. They could easily be made public
[19:56:11] <ebernhardson>	 well, "easily" as a one time thing, some process to auto-publish would of course be more work
[19:57:16] <gehel>	 Any chance of leaking pii ? My limited understanding is that it is unlikely, but... Limited understanding
[19:57:53] <ebernhardson>	 gehel: i guess we could double check with research, but the trees amount to if(title_match > 0.5) { if (near_match < 0.4) { if ...
[19:58:29] <ebernhardson>	 i think the chances are pretty low
[19:59:09] <ebernhardson>	 as a one time thing, it's mostly copying the right file out of hdfs
[19:59:28] <inflatador>	 back
[20:01:41] <ebernhardson>	 inflatador: re: beta breakage, it wasn't the search clusters. changeprop had old decomissioned kafka hostnames, and mediawiki had decomissioned eventgate hostnames
[20:06:48] <ebernhardson>	 apparently i can respond to the list from the mailman interface, nifty. Will try and write something
[20:10:08] <inflatador>	 ebernhardson gotcha, I think I remember seeing htat one now that you mention it
[20:24:03] <inflatador>	 sorry, not feeling so good. Going to pick up some meds and hopefully will see you tomorrow
[20:26:21] <ryankemper>	 inflatador: feel better!