| === chris14_ is now known as chris14 | ||
| === sdeziel_ is now known as sdeziel | ||
| paride | lvoytek, hi Lena! As far as you remember, do we still have that mysql-8.0 issue where on upgrades, if the database is big (or the machine is slow), the daemon does not shut down in time and we get "Can't start server: Bind on TCP/IP port: Address already in use" errors because of that? | 15:17 |
|---|---|---|
| paride | lvoytek, I hope I remember the issue well enough to make a bell ring. | 15:17 |
| mdeslaur | I think that was bug 1921378, I remember seeing that SRU | 15:21 |
| -ubottu:#ubuntu-server- Bug 1921378 in mysql-8.0 (Ubuntu Jammy) "mysql server crashes with port 3306 already in use" [High, Fix Released] https://launchpad.net/bugs/1921378 | 15:21 | |
| lvoytek | paride: Yeah I remember that bug. We ended up handling that by increasing the timeout value to a few minutes. Unfortunately that might not be enough for some very old systems. We could always increase the timeout more if it affects enough people though | 15:25 |
| paride | thanks lvoytek and mdeslaur, yes that's it. I don't think we have "enough people", but I have (maybe) one of those bugs to triage and I wanted to check if it actually matches that scenario. | 15:28 |
| lvoytek | paride: That's fair. I'll try and keep an eye out for those types of bugs too over time and if more pop up I'll see if it's reasonable to increase the timeout more | 15:31 |
| JanC | ideally there would be some sort of way to detect mysql is busy shutting down instead of something error-prone like a time-out... | 15:40 |
| paride | that's fundamentally how things are, the timeout is to avoid "stuck forever" | 16:16 |
| === esv_ is now known as esv | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!