[09:50:52] 10Traffic: Provide a TCP MSS clamping mechanism for real servers - https://phabricator.wikimedia.org/T350462 (10Vgutierrez) [09:51:08] 10Traffic: Provide a TCP MSS clamping mechanism for real servers - https://phabricator.wikimedia.org/T350462 (10Vgutierrez) p:05Triage→03Medium [09:52:38] 10Traffic: Provide an etcd schema to store real servers information - https://phabricator.wikimedia.org/T350463 (10Vgutierrez) [09:55:58] 10Traffic: Provide an etcd schema to store real servers information - https://phabricator.wikimedia.org/T350463 (10Vgutierrez) [10:21:01] hi! whou would be a good person to reviiew acme-chief patches like https://gerrit.wikimedia.org/r/c/operations/puppet/+/957720? [10:26:10] morning taavi :) [10:36:13] thanks! :-) [10:50:01] vgutierrez: I was thinking about testing a slight change to our router/switch BGP config facing doh2002, does that seem ok? [10:50:23] I've labbed up the change - to the TTL settings for the peering - and should be ok, but there is a slight chance it'd break the session [10:50:55] I want to test with one device and something not super important before making the change globally which would affect more important things too [11:45:24] topranks: go for it :) [11:47:17] thanks! [11:54:39] topranks: for something even less critical, there is durum as well, durum2001 and 2 [11:54:49] the wikidough check service [11:55:08] cool, I'll use those instead even [11:55:11] but I guess as long as it's not dns*, we are good! [11:55:15] * sukhe brb [12:27:12] 10netops, 10Infrastructure-Foundations, 10SRE: Netbox PuppetDB Import Script Failing for cloudnet2006 - https://phabricator.wikimedia.org/T350479 (10cmooney) p:05Triage→03Medium [13:35:17] hello! Just a little heads-up (no action required) on a change we're making to the rest-gateway. We previously enforced the Host header to be wikimedia.org for the aqs URLs to stop us from caching identical data for all sites - this worked fine for external clients, but it's a bit of a deviation for internal clients [13:35:34] so we're just *only* going to match on the wikimedia.org domain in URL paths https://gerrit.wikimedia.org/r/c/operations/deployment-charts/+/971456/ [13:36:11] behaviour as far as caches are concerned will be more or less the same, we'll 404 on different paths in the same way we would have 404'd on different Host headers [13:56:45] 10netops, 10Infrastructure-Foundations, 10SRE: Put Dell SONiC switches in production - https://phabricator.wikimedia.org/T335028 (10Jclark-ctr) [14:17:01] 10netops, 10Infrastructure-Foundations, 10SRE: Use default BGP multihop TTL between devices - https://phabricator.wikimedia.org/T350488 (10cmooney) p:05Triage→03Medium [16:42:37] 10Traffic: Provide a TCP MSS clamping mechanism for real servers - https://phabricator.wikimedia.org/T350462 (10Vgutierrez) I've explored the `BPF_PROG_TYPE_SOCK_OPS` alternative and I've detected the following caveats: 1. the eBPF program needs to be loaded before the daemon performs a `listen(2)` otherwise it... [17:12:44] (VarnishPrometheusExporterDown) firing: Varnish Exporter on instance cp4052:9331 is unreachable - https://wikitech.wikimedia.org/wiki/Prometheus#Prometheus_job_unavailable - https://grafana.wikimedia.org/d/000000304/varnish-dc-stats?viewPanel=17 - https://alerts.wikimedia.org/?q=alertname%3DVarnishPrometheusExporterDown [17:15:26] 10netops, 10Infrastructure-Foundations, 10SRE, 10Patch-For-Review: Use default BGP multihop TTL between devices - https://phabricator.wikimedia.org/T350488 (10cmooney) [18:17:50] 10netops, 10DC-Ops, 10Infrastructure-Foundations, 10ops-eqiad: eqiad: Connect IC-374549 - https://phabricator.wikimedia.org/T350504 (10RobH) [18:17:58] 10netops, 10DC-Ops, 10Infrastructure-Foundations, 10ops-eqiad: eqiad: Connect IC-374549 - https://phabricator.wikimedia.org/T350504 (10RobH)