journald is unable to attribute messages incoming from processes that exited their cgroup
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
systemd |
New
|
Unknown
|
|||
systemd (Ubuntu) |
Invalid
|
Undecided
|
Unassigned |
Bug Description
This is mostly an FYI and a tracker to link Upstream to Ubuntu.
Background:
I wondered why some of my services are missing just the most interesting "last" messages before dying.
Unfortunately I found this is a known race and there seems to be no good fix yet.
But I think this is important, so I wanted to make you aware.
Especially the last few messages before a service is dying are important.
If you see any way to fix this in Ubuntu as an interim solution until upstream has found "the right thing" to eventually solve it that would be great.
Upstream issue (many dups onto this): https:/
One approach that was tried (but not accepted): https:/
Changed in systemd: | |
status: | Unknown → New |
On the bright side messages are in journal, but not if you use -u to filter.
And if a service looks like this:
service[123]: good
service[123]: good
service[123]: <dead>
People wonder, until they look in an unfiltered journal to find the following it take some time and confusion.
service[123]: good
service[123]: good
service[123]: failing because of XYZ
service[123]: <dead>