[07:28:09] greetings [07:28:12] morning [10:18:54] XioNoX: topranks: btw, some statistics I promised: we have ~600 vms still in the legacy VLAN, compared to ~350 in the new VXLAN based network [10:25:24] andrewbogott: ok to re-enable puppet on codfw1dev cloudcontrols? it's been disabled for a while and is blocking me now [10:34:53] taavi: thanks! that's actually pretty good progress, also a good sign things on the VXLAN based stuff is working well (or we'd have heard) [10:36:43] so my thinking is it may not make sense to over-engineer the L2 vlan, we can probably live with the "manual intervention" failover we have if a switch dies now for it. And we can encourage people to use the newer networking if their service needs a better "SLA" than that. [11:58:11] neat, kubernetes is adding native support for mutating resources based on CEL rules https://github.com/kubernetes/enhancements/issues/3962 [11:58:18] (spotted while doing changelog review for 1.32) [12:01:17] very cool [13:01:47] taavi: yes, go ahead and re-enable [13:07:07] thanks, done [13:16:03] does anyone have bandwidth to look at the toolsbeta errors or should I put that on my short list? The flapping alert is driving me up the wall :) [13:36:18] I don't no [14:32:21] godog: here's moving the zookeeper cluster def into hiera, I think it's unrelated to the ID question from yesterday. https://gerrit.wikimedia.org/r/c/operations/puppet/+/1283000 [14:32:47] I sort of dislike that change because it adds one more place to update when cloudcontrols are refreshed, but it is more consistent with other zk uses. [14:34:41] andrewbogott: thank you, yeah I see what you mean about "yet another place", I think in this case being able to control when the zk cluster changes is a plus [14:35:54] ok -- I don't feel strongly, the new patch + taavi's does at least avoid a bit of regex magic. [14:36:35] indeed, I think your might need rebasing on current production? I don't see taavi's change [14:36:58] mine probably needs merging first? [14:37:47] lol indeed, rebasing on the unmerged patch [14:37:59] yeah, taavi's should go in first [14:38:47] I don't mind doing the rebase/merge/apply bits though, don't want to interrupt whatever you two are doing. [14:39:27] sure np [14:39:55] andrewbogott: unrelated, the change you suggested yesterday for cloudvps-quota https://gitlab.wikimedia.org/toolforge-repos/cloudvps-quota/-/commit/a8fbd98e8ae26c0b3502ebf92f2dedc9f0b0b931 [14:40:09] already deployed [14:40:22] nice, thank you. [14:40:42] will save me 30 seconds of confusion every time I compare the table to the cookbook command [14:42:01] indeed much clearer this way [14:43:01] Yesterday I was helping at a food pantry and one of my very few very simple jobs was subtracting the weight of an empty grocery cart from the weight of a full one, and I learned that I can absolutely NOT reliably perform subtraction in my head. Or, at least, 55 (the weight of most of the empty carts) is maybe an extra-hard number for me to subtract. [16:35:45] can I get a +1 for T425420 and T425387? [16:35:45] T425420: Increase quota for wikitextexp project - https://phabricator.wikimedia.org/T425420 [16:35:46] T425387: Quota increase for development metrics - https://phabricator.wikimedia.org/T425387 [17:40:59] dhinus: done