[07:14:32] 10SRE-tools, 10Infrastructure-Foundations, 10SRE, 10serviceops-radar: SVC DNS zonefiles and source of truth - https://phabricator.wikimedia.org/T270071 (10ayounsi) About the Ganeti, as I understand it, the issue is that they have .svc. in their fqdn, while not an SVC IP. I agree that changing the IP to so... [07:53:39] 10netbox, 10Infrastructure-Foundations, 10SRE: Evaluate Nautobot fork of Netbox and decide whether to use. - https://phabricator.wikimedia.org/T288515 (10Majavah) [08:02:31] FYI I'm grabbing sretest1001 and 1002 for a last reimage test [09:00:15] 10SRE-tools, 10Infrastructure-Foundations, 10SRE, 10serviceops-radar: SVC DNS zonefiles and source of truth - https://phabricator.wikimedia.org/T270071 (10akosiaris) >>! In T270071#7404620, @ayounsi wrote: > About the Ganeti, as I understand it, the issue is that they have .svc. in their fqdn, while not an... [10:46:06] volans: fyi i rolled out the new ipaddress6 fact . the only nodes to change where the gitlab servers which now have the correct value [10:46:21] and mx1001 I hope :D [10:48:04] yes mx now has the correct one as well [10:49:21] great [10:49:25] thanks a lot [10:51:42] np [12:09:52] 10netops, 10Infrastructure-Foundations: Eqiad Expansion - LVS Connectivity Options - https://phabricator.wikimedia.org/T292630 (10cmooney) [12:10:03] 10netops, 10Infrastructure-Foundations: Eqiad Expansion - LVS Connectivity Options - https://phabricator.wikimedia.org/T292630 (10cmooney) p:05Triage→03High [12:39:02] 10SRE-tools, 10Infrastructure-Foundations, 10SRE, 10serviceops-radar: SVC DNS zonefiles and source of truth - https://phabricator.wikimedia.org/T270071 (10fgiunchedi) As far as Prometheus goes the `svc` CNAME in PoPs is there for symmetry reasons with eqiad/codfw (where prometheus is LVS'd), so prometheus... [13:51:53] 10Puppet, 10Infrastructure-Foundations, 10GitLab (Infrastructure), 10Patch-For-Review, and 3 others: Puppetise gitlab-ansible playbook - https://phabricator.wikimedia.org/T283076 (10Jelto) I rolled out the puppetised changes to `gitlab2001` (gitlab-replica). Apart from a [minor ssl fix](https://gerrit.wiki... [14:20:13] 10netops, 10Infrastructure-Foundations, 10SRE: Eqiad Expansion - LVS Connectivity Options - https://phabricator.wikimedia.org/T292630 (10cmooney) @wiki_willy One option that Dell do seem to have are the Mellanox ConnextX-3 and ConnectX-4 QSFP+ based cards. With the right module we can in theory do 4x10G off... [14:24:13] Infra Foundations people... in relation to the above task I'm investigating options for LVS connectivity to the new Eqiad rows. [14:24:37] TL;DR is that the existing LVS don't have free PCI slots for additional NICs to connect them the same way as current rows. [14:25:01] There are some options here, but one obvious one is to use denser NICs. [14:25:22] Either 4xSFP+, or 1xQSFP+ NICs, either of which will give 4x10G off each NIC. [14:25:38] Problem is Broadcom don't seem to make such a card, the options are Intel or Mellanox. [14:25:58] I'm not sure what the feelings on those are? I know we standardize on BCM and are comfortable with the drivers etc. [14:26:08] Wondering if I should discount the possibility completely or not. [14:54:52] make sure to get also brandon's feedback on this topic ;) [14:59:12] Yep I've asked him as well. I know there is a lot of tuning etc. on the LVS [14:59:14] thanks. [15:27:49] 10netops, 10Infrastructure-Foundations, 10SRE: Eqiad Expansion - LVS Connectivity Options - https://phabricator.wikimedia.org/T292630 (10BBlack) I'm pretty sure you're right about the 2x PCIe limitation on these servers, unfortunately. What I'm not so sure about, is whether the (currently BIOS-disabled) "on... [15:36:32] 10netops, 10Infrastructure-Foundations, 10SRE: Eqiad Expansion - LVS Connectivity Options - https://phabricator.wikimedia.org/T292630 (10BBlack) Also note that lvs1013-16 are ~4 years old now. I don't think they were scheduled for refresh this year, but they probably would be next year (and by then we would... [17:17:58] 10netops, 10Infrastructure-Foundations, 10SRE: Eqiad Expansion - LVS Connectivity Options - https://phabricator.wikimedia.org/T292630 (10cmooney) Thanks for the detailed response @BBlack > And yes, the tunnel option is also risky/complex/painful, but we'll have to weigh that against all the above. Without... [19:39:57] 10netops, 10Infrastructure-Foundations, 10SRE: Eqiad Expansion - LVS Connectivity Options - https://phabricator.wikimedia.org/T292630 (10cmooney) > I know in the R720s, R730s etc the "on-board" network ports are on a daughterboard which is swappable. From everything I've seen online this isn't the case with... [21:06:01] 10netops, 10Continuous-Integration-Infrastructure, 10DC-Ops, 10ops-codfw: Flapping codfw management alarm ( contint2001.mgmt/SSH is CRITICAL ) - https://phabricator.wikimedia.org/T283582 (10Papaul) We are seeing this issue because all those hosts are running an old firmware version for the IDRAC. Upgrading...