[01:10:36] I've got an interesting problem with SSH into Cloud VPS. I'm changing the way I manage SSH keys to having a unique key per system I'm trying to access, all saved in 1Password. However, I've found that I'm now unable to SSH into bastion.wmcloud.org because I get a "Too nany authentication failures" error, [01:11:08] Based on the ssh verbose output, I think this is because it's offering the various keys I have in 1Password, but because I have so many, the server disconnects me before it can offer the correct key [01:12:36] I don't suppose anyone knows of a workaround or something? I'd prefer to keep all my SSH keys in my password manager and using its agent rather than the OpenSSH default agent [01:21:29] have you tried specifying a key in your local ssh config? [01:23:03] Hmm, can you do that with something other than the full private key? I'm aware of IdentityFile (I think?), but iirc that requires the private key file [01:28:01] https://developer.1password.com/docs/ssh/agent/advanced/#match-key-with-host [01:32:34] Oh! I was trying to find all sorts of SSH docs but I didn't think to search 1Password's [01:32:44] That's exactly what I need, thanks [01:33:22] (None of the SSH docs I did find say explicitly that you can use a public key with IdentityFile, though I guess they also don't say it must be a private key either) [01:34:21] of course this is also more efficient because you can pick the right key the first time every time. [01:34:41] Right [01:34:50] Beautiful, that worked. Thank you very much! [01:34:52] not just about working around the 6 key limit [01:35:08] sure 🎉 [05:24:37] My webservice https://hoiscript.toolforge.org won't start. [05:24:51] https://www.irccloud.com/pastebin/4QXTqwfv/shell [05:25:11] I keeps returning 503. And restarting via CLi does not work. [05:25:21] * It [05:26:36] Could someone take a look? [05:29:52] Hmmm. It returns back online now. [09:53:38] TIL `webservice logs` [10:59:40] lucaswerkmeister: yeah, fairly new thing, it reads logs from the running kubernetes pods, intended for buildservice based tools and others that don't log to NFS [11:00:07] i've been thinking if we should have some changelog page that lists minor but user-facing changes that don't get announced on cloud-announce [11:09:22] taavi: yeah, I figured it must be for something other than uwsgi.log [11:32:12] !log tools.lexeme-forms deployed fb8c4a30ff (update punjabi-noun-masculine-guru) [11:32:15] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.lexeme-forms/SAL [11:34:18] for some reason my tool seems to reliably need one more pod restart per deployment restart now [11:34:57] i.e. after each `kubectl rollout restart deployment lexeme-forms`, the new pod will be “ready 0/1” for a bit longer than usual, and when it’s eventually “ready 1/1” (and the old pod is gone), it’ll have “restarts 1 (Xs ago)” [11:35:22] doesn’t really bother me, just thought I’d mention it in case it’s relevant to anyone else [11:35:40] (maybe it’ll go away on the python3.11 image ^^) [12:50:33] !log tools.integraality Deploy 5aecae8, 649259f, 9712666 (T338684) [12:50:36] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.integraality/SAL