I have seen instances where the crash daemon is running under a container (using /var/lib/cep/{fsid}/crash), but the daemon is trying to use the legacy location (/varlib/ceph/crash). This can result in file access violations or "file not found" issues which should show up in the system logs (journalctl).

The root cause is that the container is using /var/lib/ceph/crash internallyy, but should be mounting it on the external /var/lib/ceph/{fsid}/crash. apparently it uses /var/lib/ceph as its external path to mount to, but that's not where the installation created the crash directory.

You can fake it out by creating your own /var/lib/ceph/crash infrastructure and setting the ceph userid as its owner, which is ugly, but beter than nothing.

    Tim

On 11/11/24 01:20, nguyenvand...@baoviet.com.vn wrote:
and a new issue

when we try to rm and add the crash daemon on already-running host, ceph orch 
return them

NAME                        HOST       PORTS        STATUS        REFRESHED  
AGE  VERSION    IMAGE ID      CONTAINER ID
crash.cephgw03              cephgw03                starting      -          -    <unknown> 
 <unknown>     <unknown>

before rm, its OK

I think there some problem with ceph orch
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io

Reply via email to