ceph dashboard oddities

[damo2929]

hi all

I have finished a rolling eviction upgrade and rejoin of my nodes to take them from 7.4 to 8.1 but I have an issue with the ceph dashboard as follows any ideas what could be causing it.

 

[Alwin Antreich]

Not certain, but it might be caching and vanish. How does ceph -s look? Some leftover MON/MGR/MDS?

 

[damo2929]

root@wlsc-pxmh01:~# ceph -s
cluster:
id: ab50ab95-b39b-432a-b24f-0858830fb975
health: HEALTH_OK

services:
mon: 7 daemons, quorum wlsc-pxmh10,wlsc-pxmh04,wlsc-pxmh01,wlsc-pxmh03,wlsc-pxmh05,wlsc-pxmh06,wlsc-pxmh02 (age 7d)
mgr: wlsc-pxmh01(active, since 7d), standbys: wlsc-pxmh10, wlsc-pxmh04, wlsc-pxmh02, wlsc-pxmh06, wlsc-pxmh05, wlsc-pxmh03
mds: 2/2 daemons up, 5 standby
osd: 72 osds: 72 up (since 7d), 72 in (since 7d)

data:
volumes: 2/2 healthy
pools: 23 pools, 2001 pgs
objects: 11.28M objects, 43 TiB
usage: 57 TiB used, 194 TiB / 252 TiB avail
pgs: 2001 active+clean

io:
client: 8.3 KiB/s rd, 36 MiB/s wr, 3 op/s rd, 487 op/s wr


it's only shown once on in the output. and we have no varnish etc in the path this is directly from the host web UI's for the whole cluster.

 

[Alwin Antreich]

it's only shown once on in the output. and we have no varnish etc in the path this is directly from the host web UI's for the whole cluster.

No, I meant that the browser caches this view. And (I forgot the details) there is also some kind of state broadcast of those services between the nodes. I believe this in the pvestatd daemon, maybe restarting this daemon on the nodes might help clear that up.