[09:11:21] dcausse: are you around [12:54:45] greetings! [13:03:42] Just pinged data persistence in their chat room, hopefully we can get https://phabricator.wikimedia.org/T309715 taken care of soon [13:46:00] dropping off my sons, back in ~15 [14:02:57] and back [14:03:28] looks like we have creds to the thanos-swift cluster now, gonna give those a shot [15:03:01] If anyone has time to look at https://gerrit.wikimedia.org/r/c/operations/puppet/+/803300/ LMK, I don't understand why jenkins doesn't like it [15:03:40] it warns "double quoted string containing no variables (double_quoted_strings)" but I don't get why that's a failure [15:05:13] looks the same as other file_line invocations, at least as far as I can tell [15:05:31] inflatador: if there are no variables (like f-strings in python) you need to use single quotes [15:05:41] it's kinda like you had an f-string without any variable replacement inside [15:07:06] inflatador: triage? https://meet.google.com/eki-rafx-cxi?authuser=1 [15:07:27] OMW [15:07:29] https://puppet.com/docs/puppet/7/lang_data_string.html#lang_data_string_single_quoted_strings [15:07:57] actually, to be precise [15:07:57] https://puppet.com/docs/puppet/5.5/lang_data_string.html#single-quoted-strings [15:08:57] and to be clear, it's a style thing [15:12:40] volans Ah OK, got it...it's interpreting the line as a puppet var, I guess I need to escape it [15:14:11] inflatador: no just use single quotes [15:14:19] when not using any variable for interpolation [15:14:45] Will give it a try, thanks volans [15:20:51] LOL, now it thinks it's a variable and I should be using double quotes [15:21:44] which one? [15:22:11] ah ${PATH} [15:22:43] line 145, yeah [15:22:58] I guess I can just go with "$PATH" [15:23:21] so for that use "PATH=\${PATH}...." [15:25:38] ACK [15:25:38] see https://puppet.com/docs/puppet/5.5/lang_data_string.html#escape-sequences [16:10:12] OK, Jenkins is happy, anyone have time for review? https://gerrit.wikimedia.org/r/c/operations/puppet/+/803300 [16:11:08] inflatador: seems easy enough, +1 [16:11:27] muchas gracias ebernhardson [17:08:27] ebernhardson or anyone else, do you know if I can find the thanos-swift creds for flink on the wdqs servers? I'm not having any luck logging in with our new swift thanos user, and I want to compare how it's set up for flink, specifically the api key value at link 28 here https://gerrit.wikimedia.org/r/c/operations/deployment-charts/+/769699/2/charts/flink-session-cluster/templates/_flink-conf.tpl#b27 [17:10:18] hmm, not sure but lets look :) I suppose i would guess somehow those come from helm and get injected into perhaps a mounted volume of the container? [17:10:46] yeah, I think you're right. Was hoping to get lucky but I don't see it on wdqs [17:11:21] if you can login to the host running the container you should be able to run a new container that mounts the volume and reports its contents (would have to check how exactly), otherwise can figure out where helm is getting its data [17:12:05] yeah, I think I can do that...also the helm secrets might be on puppet, I think I did a patch for that recently [17:12:45] openstack (or at least the weird one we used to run at Rackspace) had separate concepts of API key and password. I don't think there is a distinction in vanilla openstack, but I want to make sure I'm not missing anything [17:13:55] looking at the volume should be something like `docker run --rm -i -v=:/mnt --entrypoint bash /bin/bash find /mnt` [17:13:57] (untested) [17:14:15] err, probably only invoke bash once [17:18:00] this works locally at least: docker run --rm -it -v=mwcli-mwdd-default_elasticsearch-data:/mnt --entrypoint /bin/bash docker-registry.wikimedia.org/releng/ci-stretch:latest -c "find /mnt" [17:20:52] I'll dig around a bit, I was hoping to find the info on the deploy server, but it's just templates so far. I guess the secrets are probably in etcd. Maybe peeking in the container is easier [17:23:24] i see reference to some /etc/helmfile-defualts/private, but sadly i know nothing about where it comes from or what it is [17:40:07] I notice the swift usernames are $account:$user as opposed to $user in some places, like hieradata/common/profile/swift.yaml [17:40:13] maybe that's the problem [17:47:21] yeah, that was it. Needed to add a colon as in the PR https://gerrit.wikimedia.org/r/c/operations/puppet/+/803284/2/hieradata/common/profile/thanos/swift.yaml [17:48:01] OK, so now we can presumably get uploading! (After I eat lunch, that is!)...back in ~45 [18:25:09] back [18:52:31] ryankemper ebernhardson this is the tentative plan for Ryan and I's pairing session today at 2 PM PDT . Feel free to add/change/respond here or in the ticket https://phabricator.wikimedia.org/T309648#7983820 [18:54:01] I imagine we'll see some weirdness with elasticsearch-keystore and our multiple instances, but I'm guessing we can work around with a symlink [18:56:24] dr appointment, back in ~1h [19:02:07] meh, mwdd detects --port=4444 as "Please set environment var --port to 4444 [19:02:29] (instead of passing along as a command arg) [19:58:29] back [19:58:51] Also that's gross CLI behavior ;P [20:01:11] yeah, I think I'd prefer it accepting either and going with the actual arg if both are present [20:07:48] i suppose i should patch it, it doesn't seem intentional. They take any arg matching '\w+=\w+' (without anchors, afaict) [21:08:16] ryankemper ebernhardson up at meet.google.com/yez-bmya-hge if you wanna join, although strugglling with swift ATM [21:29:58] off a bit early, taking liam to ortho