-
Notifications
You must be signed in to change notification settings - Fork 7
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
>=raft-0.18.2: "test_uv_truncate_snapshot.c:42: assertion failed: status == result->status (23 == 0)" #35
Comments
With Fedora 38 I can reproduce this issue with every build (only tried x86_64 so far). However with Fedora 39 and Rawhide I cannot reproduce it and the test suite passes. They all use the same |
This one is probably spurious. The test suite assumes that the host system is not running workloads that could interfer with the test, for example processes that use the AIO kernel subsystem. Is the system running something like mysql, postgresql or other similar service that might make use of asynchronous I/O? |
This one might be a similar reason as the other one, error code Are these tests being run in parallel? (e.g. with |
If there's no way for you to run these tests in a fully isolated environment, one thing I could do is to disable this kind of environment-sensitive tests by default, and only run them if some special environment variable is set (which we would do in our own CI). |
I'm test building this on my workstation via Let me submit it to COPR and see if it works better there. But as far as I understand the Fedora build infrastructure is on a shared kernel using mock too.
Yes, it's run with |
I see. I think I'll disable those tests by default then.
Ok, please try to run that without |
Ok, the COPR environment seems to be more to the liking of the test suite 😃 It built successfully for all architectures and Fedora releases: https://copr.fedorainfracloud.org/coprs/ganto/lxc4/build/6781882/ (c875d31) |
Great :) In that case, I'd rather not put any workaround in place in the upstream code, at least for now. What do you think? |
Yes, I agree. I'll also try to build it on Koji (the official Fedora build system) later to check if anything pops up there. I'll let you know. |
@ganto This looks related to a raft conflict I got while |
That was expected as I haven't rebuilt dqlite against the new raft version. Should be fixed now. |
@freeekanayaka: Unfortunately the official Fedora build service Koji seems to be a bit more sensitive:
A way to only selectively run these tests would be helpful in my case. |
@ganto I had prepared cowsql/raft#109 which I could sure merge. However, I'd like to first rule out the possibility that these failures are caused by those two tests being run in parallel. Please can you confirm that this build is not running these tests with If it is running them with If it's not possible or easy for you to drop the |
Ok. After the recent success I completely forgot about that. Will need to dig into the RPM macros to see how this works. |
With serial |
Attempting to build the RPM of
>=raft-0.18.2
(from cowsql/raft) via mock on Fedora 38 fails with:The text was updated successfully, but these errors were encountered: