Em dom., 1 de dez. de 2019 às 09:05, Erdem Agaoglu
<erdem.agaoglu(a)gmail.com> escreveu:
Hi Rodrigo,
The fact that you're getting logs from mon and the function name set_mon_vals
suggests that you made use of mon provided centralized config, like `ceph config set osd
cluster_network x.x.x.x` but it seems cluster_network and public_network cannot be
centralized and should be set in ceph.conf. (Just an idea not really an expert).
If that's the case you should be able to `ceph config rm osd cluster_network` to get
your cluster healthy again. Then you could try setting in config file directly.
Unfortunately this wasn't the case. I say so because to get my cluster
healthy again I just commented out the cluster_network settings in my
ceph.conf files. After that I could restart my OSD with any mon
errors. I did that a few days ago and forgot to post this info here.
I'm sorry.
My immediate issue was solved but my main issue remains: how can I
migrate my current single network board setup to a dual one?
As far as my experience goes, setting cluster_network in the global
section of my ceph.conf files isn' t the proper way.
Maybe cluster_network shouldn' t be set in the global section. Is that
it? In which section should it be set instead?
Regards,
Rodrigo Severo