[07:19:28] hello folks, two updates: [07:19:54] 1) varnishkafka on cp4037 is running a new PKI cert, if you see any alarm or weird behavior please ping me [07:20:36] 2) the retention in druid for webrequest live data is now 30 days, so in a couple of weeks we should have an entire month of live data (and we may be able to deprecate the 128 datasource if we want) [09:28:20] mac users: is dos2unix installed by default and if so does it have the -i / --info flag? [09:28:38] jbond: it isn't [09:28:44] hnowlan: thanks [12:31:29] cdanis: JameelKaisar: Just noticed your status code fix, I don't think null is possible there after calling Math.round(). I suspect what's happening is that this is NaN at that point, and it becomes null or undefined after JSON serialisation since NaN is not possible in JSON. [12:32:04] There's a few ways to work around this which I'm sure you'll enjoy figuring out :) [12:32:12] Hope that helps! [13:11:03] thanks Krinkle, that makes sense [13:21:56] Hi folks, puppet isn't working on the swift frontends (at least) any more - error is https://phabricator.wikimedia.org/P49422 [13:22:07] Krinkle: cdanis just realised this shouldn't have worked even if Math.random() returned null. I forgot to remove status_code on line 299 🥲 [13:22:21] I've not made a change in that area recently... [13:23:09] Emperor: from a quick glance at https://gerrit.wikimedia.org/r/plugins/gitiles/operations/puppet/+log/refs/heads/production [13:23:12] https://gerrit.wikimedia.org/r/plugins/gitiles/operations/puppet/+/cced2208749e6be7cf3f246ccbfc401e74ccdb46 [13:23:20] Yeah, I'm just looking through git log [13:23:20] Emperor: I think it was likely https://gerrit.wikimedia.org/r/c/operations/puppet/+/929383 [13:23:30] yeah, same [13:23:30] at least, that has that class name in the description and says 'drop' :) [13:23:39] but correlation might be causation :) [13:24:03] if jbond isn't around I suggest just reverting the patch for now [13:24:37] jbond: are you around? I think tlsproxy::localssl: wasn't as unused as you thought... [13:25:18] * Emperor will revert shortly if no response [13:26:04] sorry im here [13:26:15] you can revert the tlsproxy patch if its still in use [13:26:22] * jbond just reading backlog [13:26:45] jbond: thanks, will do imminently. its in use on swift and maps at least AFAICT from git grep [13:27:04] Emperor: please revert i must have had a brain fart this morning its definetly still in use [13:27:09] doing so [13:27:20] sorry for the interuption [13:27:46] https://gerrit.wikimedia.org/r/c/operations/puppet/+/929650 if you fancy a +1 :) [13:29:47] ugh, CI failure [13:30:49] no SPDX header :( [13:30:51] * jbond looking [13:31:23] Emperor: i have overriden ci feel free to submit and merge [13:31:32] TY [13:34:05] Emperor: i also re-opened the task. and taged maps & swift. is possible the original issue is not a problem for those so feel free to close it again [13:34:16] in /win 14 [13:38:32] jbond: if I understand correctly the failure mode is "puppet always flags this resource as changing on every puppet run"? In which case we don't have that problem on swift (e.g. the most recen ms-fe1009 run which is Unchanged) [13:51:36] Emperor: ack i closed it again [13:59:20] thanks! [14:00:58] JameelKaisar: ah, good call, I hadn't noticed! Yeah, it's still in the object! [14:07:51] hello service-ops anyone willing to help me with this https://phabricator.wikimedia.org/T326564 this Thursady morning 10am CT thanks [14:11:21] That's 1500 UTC right ? [14:13:37] papaul: if my time conversion's right I can help. Depool, downtime, shutdown ? [14:15:19] claime: yes thanks [14:15:33] depool downtime and shutdown [14:16:29] ack [17:53:49] Can anyone help me understand the CI failure here? https://gerrit.wikimedia.org/r/c/operations/cookbooks/+/914018/ ? Error msg is `cookbooks/sre/wdqs/data-transfer.py:122: error: "Sequence[str]" has no attribute "append" [attr-defined]` . The code itself works [17:58:25] inflatador: the type checker doesn't like that you've used the same name `instance` at line 111 and line 119, for different types [17:59:22] so at 111 it deduced the type of `instance` to be whatever it is in that context, and then at line 122 you did something that wouldn't be valid for that deduced type [18:00:12] the quickest fix is just to use two different variable names in those two different if/elif blocks, rather than calling them both `instance` [18:01:48] rzl thanks, will try changing the var name [18:18:36] hmm, actually I do need 'instance' to be the same for both...maybe I'll try breaking out into a separate function