Logs are hammered with a healtheck error in empty /var/log/containers/stdouts/healthcheck.log
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tripleo |
New
|
Undecided
|
Unassigned |
Bug Description
On undercloud node, /var/log/messages is overflowing with messages like these:
Nov 2 00:00:13 undercloud podman[818647]: 2020-11-02 00:00:13.013180118 -0500 EST m=+0.892513654 container exec 4218e223289e3dd
Nov 2 00:00:13 undercloud podman[818629]: healthy
Nov 2 00:00:13 undercloud podman[818629]: Error: unable to update health check log /var/log/
Nov 2 00:00:13 undercloud podman[818629]: , error found in #0 byte of ...||..., bigger context ...||...
Nov 2 00:00:13 undercloud systemd[1]: 2849cfe4117b4eb
Nov 2 00:00:13 undercloud systemd[1]: 2849cfe4117b4eb
Nov 2 00:00:13 undercloud podman[818647]: healthy
Nov 2 00:00:13 undercloud podman[818647]: Error: unable to update health check log /var/log/
Nov 2 00:00:13 undercloud podman[818647]: , error found in #0 byte of ...||..., bigger context ...||...
The rate of these messages is several such blocks EACH SECOND.
They can be suppressed for a short period of time by doing
echo '{}' > /var/log/
but they soon begin anew as something truncates the file.
This is happening with the trunk tripleo, installed with
"tripleo-repos current" (with no -b of any kind).
I am unable even to figure out which component is responsible for parsing the JSON
in the log, and/or which component truncates the file repeatedly.