Dear Stefan,
thanks, I tried your suggestion. Unfortunately no matter whether I put
mon_initial_members = hostname1,hostname2,hostname3
or
mon_initial_members = a,b,c
into ceph.conf (both on deployment and mon host),
"ceph-deploy gatherkeys" still tries to use
"--admin-daemon=/var/run/ceph/ceph-mon.ceph-node10.asok"
instead of "--admin-daemon=/var/run/ceph/ceph-mon.a.asok".
When I create "/var/run/ceph/ceph-mon.ceph-node10.asok"
on the node (as a symlink), I get another error instead, one
step later:
[ceph-node10][ERROR ] could not find 'MON_HOSTNAME' in monmap
Any ideas?
Cheers,
Oliver
On 04/20/2018 11:46 AM, Stefan Kooman wrote:
Quoting Oliver Schulz (oliver.schulz@xxxxxxxxxxxxxx):
Dear Ceph Experts,
I'm try to switch an old Ceph cluster from manual administration to
ceph-deploy, but I'm running into the following error:
# ceph-deploy gatherkeys HOSTNAME
[HOSTNAME][INFO ] Running command: /usr/bin/ceph --connect-timeout=25 --cluster=ceph --admin-daemon=/var/run/ceph/ceph-mon.HOSTNAME.asok mon_status
[HOSTNAME][ERROR ] "ceph mon_status HOSTNAME" returned 22
[HOSTNAME][DEBUG ] admin_socket: exception getting command descriptions: [Errno 2] No such file or directory
The problem seems to stem from the fact that the cluster still uses
the naming scheme "mon.a", "mon.b", etc., while ceph-deploy assumes
"mon.HOSTNAME".
Is the a workaround for this, resp. a way to fix the cluster? Or
will it be necessary to recreate all mons?
I believe it might have to do with the requirement for "ceph-deploy" to
have a "mon_initial_members = your MONs here". Do you have that in your
ceph.conf file on the node you are deploying from?
Gr. Stefan
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
