I am trying to get detailed information about the RBD images used by
OpenStack (r/w operations, throughput, ..).
On the mailing list I found instructions that this is possible using an
admin socket of the client . So I enabled the socket on one of my
hosts according to . The manual states that the socket should be
there once I restart the VM. At some point it actually does appear, but
it vanishes within a second or two. If I keep monitoring the directory I
see it appearing for roughly 1-2 seconds per minute.
The socket looks like this:
root@compute01:/var/run/ceph/guests# ls -l
srwxr-xr-x 1 cinder cinder 0 Aug 29 17:54
Does anyone know what I am doing wrong?
Or is there another way to get information about which RBD image is
causing the most load on a cluster?
By the way, the mail archive from before 2019 seems to be inaccessible.
Using google cache as a fallback.