Don't coredump on unavailable postgres #57
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
To reproduce:
This probably happens because of Lix's crash handler doing an abort(); to get a coredump.
This is a little harsh for a failing DB connection. It should be logged and failed (ideally, after some retries).
I was about to open an issue for hydra-queue-runner coredumping randomly (I assume that postgres runs out of connections or something), so it might be that.
If someone manages to produce a fix for this, please ping me so I can check if this is my problem, or I'm having a different one.
ginormous stacktrace just in case
btw, finally managed to update to NixOS 25.05 Hydra
0.1.19700101.e8f30d6
(usinglix-2.94.0-dev-2d01098
), so I'm a viable test candidate again as soon as I've dealt with the fallout from that. At that point I'll probably snag the other PR onto my box.At first glance that seems like a different issue. Do you get something in the logs before the crash by any chance?
Ah, yeah, nevermind. The error and the coredump were quite far apart in some cases. I'm running into Too many open files consistently.
I guess it's kind of the same problem (lack of error handling), and the solution similar to the postgres connection would be literally anything other than coredumping, though it may be different enough to warrant a separate issue.
Give this post a thumbs up or something and I'll go open that separate issue.