rasdaemon process log take all the root path space

Null

An unknown issue happen when create a VM on the UIS.

Migration from other nodes to it failed and also copying files to the destination device failed.

1.    First, SSH into the problematic host and run `ceph -s`. and find that the Ceph status is HEALTH_WARN, with 1/3 mons down. The alarmed device is the problematic host.

2.    Check the process of ceph_mon, and found that this process keep restart.

3.    Check the partition status using df -h and found that the root path is full.

4.    Use du -sh * | grep G to search for large files level by level. Found the file /var/lib/rasdaemon/ras-mc_event.db, which occupies 88G of space.

5. The composition of the file

The file is the log process for the rasdaemon process. You can stop the process first and then use rm to delete it. The log file will be recreated when the process starts next time.

Steps to stop:

1) systemctl stop rasdaemon

2) rm /var/lib/rasdaemon/ras-mc_event.db

3) systemctl start rasdaemon

Temporary solution:

You can first stop the process and then directly use
        1) systemctl stop rasdaemon

        2) rm /var/lib/rasdaemon/ras-mc_event.db

        3) systemctl start rasdaemon

Solution:

Please call TSC.