Bug #46252
closed
MGRs should get a random identifier, ONLY if we're co-locating MGRs on the same host
Added by Sebastian Wagner almost 4 years ago.
Updated about 2 years ago.
Description
admin: # ceph -s
cluster:
id: 40c6ac7e-b48c-11ea-ac77-52540020a5c7
health: HEALTH_WARN
Reduced data availability: 1 pg inactive
OSD count 0 < osd_pool_default_size 3
services:
mon: 1 daemons, quorum ses7-01 (age 22m)
mgr: host-01.jixzru(active, since 22m)
osd: 0 osds: 0 up, 0 in
We probably should add the jixzru suffix, only if we're co-locating MGRs on the same host.
- Priority changed from Normal to Low
- Subject changed from MGRs should get a random identifier, if we're co-locating MGRs on the same host to MGRs should get a random identifier, ONLY if we're co-locating MGRs on the same host
is there any practical issue of having the suffix on the mgr daemon?
Redouane Kachach Elhichou wrote:
is there any practical issue of having the suffix on the mgr daemon?
No issue at all. It's just not required.
In that case (since it's not affecting the correct functioning of mgr) maybe it's better to keep consistency (using prefix always) instead of adding logic to handle colo cases separately. If you agree we can close this ticket.
- Assignee set to Redouane Kachach Elhichou
Redouane Kachach Elhichou wrote:
If you agree we can close this ticket.
Please, don't wait for my consent. I'm just here to clarify things.
great, thanks. Closing this ticket.
- Status changed from New to Won't Fix
- Status changed from Won't Fix to Closed
Also available in: Atom
PDF