[06:10:18] 10Traffic, 10Analytics, 10Analytics-Kanban: Review use of realloc in varnishkafka - https://phabricator.wikimedia.org/T287561 (10elukey) @odimitrijevic the author of the patch contributed to Varnishkafka before, they seem to know the codebase but the specific realloc patch seems to target a use case that we... [08:41:06] 10netops, 10Data-Persistence-Backup, 10Infrastructure-Foundations, 10SRE, 10bacula: Understand (and mitigate) the backup speed differences between backup1002->backup2002 and backup2002->backup1002 - https://phabricator.wikimedia.org/T274234 (10jcrespo) I see a huge improvement on the "stability" (if you... [10:53:54] 10Traffic, 10MW-on-K8s, 10Performance-Team, 10Release-Engineering-Team, and 2 others: Serve production traffic via Kubernetes - https://phabricator.wikimedia.org/T290536 (10jijiki) >>! In T290536#7364817, @Joe wrote: > I have some alternative ideas. Specifically, right now we have a limited number of diffe... [14:00:51] 10HTTPS, 10Wikidata, 10Wikidata-Campsite: Fix broken https at https://query.commons.wikimedia.org/ - https://phabricator.wikimedia.org/T291542 (10Bugreporter) [16:12:47] 10Traffic, 10MW-on-K8s, 10Performance-Team, 10SRE, and 2 others: Serve production traffic via Kubernetes - https://phabricator.wikimedia.org/T290536 (10thcipriani) [17:14:12] 10Traffic, 10SRE, 10MW-1.35-notes (1.35.0-wmf.40; 2020-07-07), 10Patch-For-Review, and 2 others: Harmonise the identification of requests across our stack - https://phabricator.wikimedia.org/T201409 (10Krinkle) [19:02:01] hey traffic, I'd like to merge/deploy part 1 of the wcqs lvs work today: https://gerrit.wikimedia.org/r/c/operations/puppet/+/713959 any gotchas I should look out for? [19:04:21] also a more specific question: right now we have the state set as `service_setup` in that patch, but I see that when it's time to configure the load balancers we want to set it to `lvs_setup`. does it sound right to merge & deploy first with `service_setup`, and then push a patch to switch the state to `lvs_setup` and nothing else, and then go through the deploy steps here? [19:04:22] https://wikitech.wikimedia.org/wiki/LVS#Configure_the_load_balancers [19:52:25] One more quick question about https://wikitech.wikimedia.org/wiki/LVS#Create_an_entry_in_the_service::catalog. It says that: [19:52:27] > Here we've defined state to be "service_setup"; this means that this service will not be included in monitoring, LVS configuration, or DNS Discovery at the moment. Until you perform the next step in the procedure, adding this stanza will be a no-op. [19:53:11] Is the `next step` it's referring to https://wikitech.wikimedia.org/wiki/LVS#Add_the_IPs_on_the_backend_servers, which is the immediate next section in those docs, or is it referring to https://wikitech.wikimedia.org/wiki/LVS#Configure_the_load_balancers, which is where `service_setup` actually gets changed to `lvs_setup`? [20:04:57] For my first question about whether it makes sense to merge first as `service_setup` then do `lvs_setup` later, it sounds like that's the right way to do things based off this state machine diagram: https://upload.wikimedia.org/wikipedia/labs/b/bf/Lvs_state.png - and presumably means since we're not doing the pybal restart-y stuff that it's safe to merge [20:36:37] ryankemper: if it's helpful, https://gerrit.wikimedia.org/r/q/topic:shellbox2-lvs is how I staged my patches, creating one for each step [20:50:22] legoktm: thanks, very helpful! yeah I think one patch per `service::catalog.$SERVICE.state` change makes a lot of sense [20:53:37] the first service_setup patch is a no-op on LVSes, but might make changes elsewhere, like for k8s services it makes some changes on the k8s worker nodes