[01:20:54] !log tools.pbbot Deploy 465b0a8: support new and legacy Toolforge URL schemes [01:20:57] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.pbbot/SAL [12:54:38] hi [12:54:48] so I did the steps here: https://tools.wmflabs.org/admin/ [12:54:57] including https://toolsadmin.wikimedia.org/tools/membership/status/795 [12:55:08] "Status: Approved" [12:55:20] Now if I click on: https://toolsadmin.wikimedia.org/tools/create [12:55:33] I get "You are not currently authorized to view the URI /tools/create." [12:57:46] CustosLimen: log out and back in [12:58:31] ah, thanks [13:02:29] So I'm trying to get this to work: https://www.mediawiki.org/wiki/OAuth/For_Developers#OAuth_2 [13:02:35] I dunno if this is right place to ask [13:02:38] but I dunno where else [13:02:55] But it does not work: https://www.mediawiki.org/wiki/Topic:Vmwiwu073xo33eh5 [13:03:08] "Application Connection Error: Client authentication failed (e.g., unknown client, no client authentication included, or unsupported authentication method)" [17:45:47] !log tools.indic-wscontest Migrate indic-wscontest to new toolforge URL (T253456) [17:45:50] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.indic-wscontest/SAL [17:45:50] T253456: Migrate indic-wscontest to new toolforge URL - https://phabricator.wikimedia.org/T253456 [21:25:34] tools-sgebastion-07 is very slow [21:30:44] looking [21:31:41] load avg is around 10 [21:36:12] doesn't look like anything particularly CPU heavy is going on [21:36:46] lots of iowait sometimes [21:38:24] forgotten exactly what I'm looking at but nfsiostat looks... ok... [21:40:15] I am trying to edit a small file in vim, and it takes a long time only to open the file [21:44:04] something's up with this VM and it's not immediately obvious to me what [21:48:17] looks like whatever this is started around 1845 [21:48:38] am looking at https://grafana-labs.wikimedia.org/d/7fTGpvsWz/toolforge-vm-details?orgId=1&var-VM=tools-sgebastion-07 [21:58:58] er, 1845 is BST [21:59:07] https://prometheus.wmflabs.org/cloud/graph?g0.range_input=6h&g0.expr=node_load1%7Binstance%3D%22tools-sgebastion-07%22%7D&g0.tab=0 shows 1745 UTC [22:01:58] now it is fast again [22:02:27] yeah whatever was going on has stopped [22:05:56] looking at the graph, this has happened earlier for a while but then it resumed [22:47:21] So I am on debian 9.6 and trying to do an apt-get upgrade and it's asking me to downgrade a lot of packages that are forcing 404s [22:48:19] 47 packages it wants to downgrade [23:17:51] also is there an automatic backup of cloud VPS instances, or is that our responsibility? [23:22:56] AmandaNP, that's your responsibility [23:23:07] ok, fair enough [23:23:12] AmandaNP, what instance is this? [23:23:28] it was utrs-database2 [23:27:16] it looks like your package downgrades (or at least the couple I checked) relate to a special wikimedia apt component - jessie-wikimedia/openstack-mitaka-jessie [23:28:23] not sure if that component's existence still makes sense actually [23:28:36] heh [23:28:47] ooh [23:28:50] apt-get update fails [23:28:56] Err:6 http://apt.wikimedia.org/wikimedia jessie-wikimedia InRelease [23:28:56] The following signatures were invalid: DB3DC2BD4CD504EF2D908FC509DBD9F93F6CD44A [23:34:30] ah yes removed in https://gerrit.wikimedia.org/r/c/operations/puppet/+/593223 [23:35:42] and removed from clients puppet catalogs in https://gerrit.wikimedia.org/r/c/operations/puppet/+/585814/1/modules/openstack/manifests/commonpackages/vms/mitaka.pp - but not absented [23:39:25] looks like /etc/apt/sources.list.d/openstack-mitaka-jessie.list still exists across 106 cloud VPS instances [23:41:04] !log utrs Manually removed /etc/apt/sources.list.d/openstack-mitaka-jessie.list on utrs-database2 to apply https://gerrit.wikimedia.org/r/c/operations/puppet/+/585814/1/modules/openstack/manifests/commonpackages/vms/mitaka.pp (missing absent) [23:41:06] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Utrs/SAL [23:41:16] AmandaNP, apt-get upgrade should look better now [23:41:37] ok, if this happens on another instance, what can I do that you just did? [23:41:54] yeah [23:42:28] might also happen on utrs-beta-production and utrs-production2 [23:43:06] yep, i'm just wondering what you did though. I'm not too familiar with the game of apt [23:43:40] just removed one of its sources that probably should not have been there for a while [23:44:39] I think the component was there to bring jessie-like packages forward into stretch instances and pin them at a higher priority than the stretch ones, providing compatibility with the older OpenStack Mitaka APIs that were in use until more recently [23:44:43] or something along those lines [23:45:31] I think some code that runs on all wikimedia puppet clients relies on connecting to the OpenStack APIs so it ends up with a bunch of dependencies like this