2025-08-24T04:21:13  *** teepee_ is now known as teepee
2025-08-24T13:21:27  *** teepee_ is now known as teepee
2025-08-24T16:29:19  <acidsys> cboltz: can you confirm my observation in poo#187719, I did a few iterations and I think my use off ssh connection reuse sometimes does not get me reliable results
2025-08-24T17:13:06  <cboltz> indeed, ssh to Leap VMs takes a very long time - for example, it took 64 seconds for "ssh water.i.o.o date" (while Tumbleweed machines work as expected)
2025-08-24T17:14:34  <cboltz> session reuse is indeed relevant - opening an additional ssh connection is as fast as I'd expect
2025-08-24T17:17:10  <acidsys> okay thanks for testing, so then it's not just me
2025-08-24T18:47:34  <cboltz> what's wrong with galera2 and galera3? monitoring complains, and I can't ping them
2025-08-24T18:47:55  <cboltz> interestingly services using galera (like the wiki) are still alive, so they can't be completely dead
2025-08-24T18:56:21  <acidsys> salt -N falkor virt.list_active_vms only shows galera1, need to check but takes some time with the coffee-break feature in new kanidm
2025-08-24T19:15:15  <acidsys> apparently falkor20 was marked for fencing because monitor run into a timeout
2025-08-24T19:16:02  <acidsys> now it does not log why it run into a timeout, but the first event happened three minutes after updating kanidm. so I would guess that whatever is making ssh slow is also making some other resolution (too) slow
2025-08-24T19:16:37  <acidsys> I downgraded it on thor + falkor* for now
2025-08-24T19:19:19  <cboltz> thanks!
2025-08-24T19:19:34  <cboltz> unfortunately monitoring still thinks that galera[23] are down
2025-08-24T19:20:05  <acidsys> the vms are online again, but mariadbd still needs recovering, I am working on it
2025-08-24T19:25:18  <cboltz> ok, thanks!
2025-08-24T19:30:05  <acidsys> is done (didn't need to do anything manual, just verified)
2025-08-24T19:30:20  <cboltz> :-)
2025-08-24T21:11:39  *** teepee_ is now known as teepee