[10:42:23] 10SRE-tools, 06Infrastructure-Foundations, 06Data-Platform-SRE (2026-03-27 - 2026-04-17): debmonitor-client crashes for growthbook image - https://phabricator.wikimedia.org/T423413#11828883 (10brouberol) 05Open→03In progress [10:42:24] 10SRE-tools, 06Infrastructure-Foundations, 06Data-Platform-SRE (2026-03-27 - 2026-04-17): debmonitor-client crashes for growthbook image - https://phabricator.wikimedia.org/T423413#11828882 (10brouberol) [11:03:38] 07Puppet: Add PATCH method to Wmflib::HTTP::Method - https://phabricator.wikimedia.org/T392096#11828941 (10Fabfur) 05Open→03Resolved Thanks, this should've been closed long long time ago... [11:21:56] FIRING: MaxConntrack: Elevated conntrack usage on krb1002:9100 - https://wikitech.wikimedia.org/wiki/Monitoring/check_conntrack - https://grafana.wikimedia.org/d/oITUqwKIk/netfilter-connection-tracking - https://alerts.wikimedia.org/?q=alertname%3DMaxConntrack [11:26:56] RESOLVED: MaxConntrack: Elevated conntrack usage on krb1002:9100 - https://wikitech.wikimedia.org/wiki/Monitoring/check_conntrack - https://grafana.wikimedia.org/d/oITUqwKIk/netfilter-connection-tracking - https://alerts.wikimedia.org/?q=alertname%3DMaxConntrack [15:00:43] 10netops, 06Infrastructure-Foundations: codfw: upgrade routers (2026) - https://phabricator.wikimedia.org/T417871#11830203 (10ops-monitoring-bot) Icinga downtime and Alertmanager silence (ID=2d6251c1-af29-449c-8ada-37bb25d75cdf) set by root@cumin2002 for 1:00:00 on 3 host(s) and their services with reason: rou... [15:42:02] elukey: jasmine and i are looking at the wikikube-ctrl hosts that were broken during switch from BIOS to UEFI, and I'm wondering if they actually need to be reimaged following the provision cookbook run [15:44:29] claime: hey, it depends - if you moved them to UEFI yes, otherwise in theory no [15:44:43] I don't recall what happened, why they were broken etc.. [15:50:03] elukey: yeah that's the issue, they were switched from BIOS to UEFI and never reimaged [15:57:08] claime: yeah, and remember to change the partman settings to include the uefi partition before reimage [15:57:14] they had the "4" bug [15:57:22] so they need be reimaged as bios [15:57:25] ah lovely [15:57:27] then re-imaged as uefi [15:57:38] provisioned as bios then uefi right? [15:57:42] right [15:58:20] claime: ok so you'd need to run provisioning first with the --legacy option, then the same without it to force UEFI, and finally reimage [15:58:35] that should circumvent the bug that you faced while provisioning (thanks to SUpermicro) [15:58:58] claime: our previous provisioning code exposed a bug in supermicro's firmware, when it switches from bios to uefi mode, we fixed the bug for future reimages, but this dance is necessary for these hosts [15:59:27] if the dance doesn't work, please let me know [16:03:45] So they were originally imaged as bios [16:03:51] then changed to UEFI but not reimaged again [16:04:20] So we should only reimage them once now and that should be good right? [16:04:56] Or do we need to reswitch them to BIOS using the --legacy option, then reimage, then provision for UEFI, and reimage once more? [16:34:48] (oops, that is in fact highlighted in your email correctly jhathaway but I mistakenly assumed it was isolated sequence for testing purposes) [16:34:48] In that case, I will dual provision + re image and confirm if it's booting up correctly, thanks! [16:36:52] claime and jasmine_: 1) provision as bios 2) provision as uefi 3) reimage [16:37:02] only one reimage should be necessary [16:37:49] jhathaway: Thanks <3 [16:37:58] ty! [16:47:55] 10netops, 06Infrastructure-Foundations: codfw: upgrade routers (2026) - https://phabricator.wikimedia.org/T417871#11830767 (10Papaul) [16:48:30] 10netops, 06DC-Ops, 06Infrastructure-Foundations, 10ops-eqiad, 06SRE: Standardize management routers interfaces - https://phabricator.wikimedia.org/T421674#11830771 (10Papaul)