Description of problem: Currently the default OSD dir is set to /srv/data while Ceph expect all OSDs to be in /var/lib/ceph/osd. This isn't a big issue since Ceph will automatically create symlinks to /var/lib/ceph/osd if the OSDs are mounted elsewhere, but if there are no technical reason for putting OSDs in /srv/data, we should leave it default.
This bug did not make the OSP 8.0 release. It is being deferred to OSP 10.
Hi Angus, I could try to drive this change via upstream getting TripleO providing sensible defaults to the deployment. If you are ok with that feel free to assign the bug to me and I'll start working on it.
this is set in a single file [1] but we'll need to figure what to do on upgrades one approach could be to prevent puppet-ceph from changing these values on a running cluster another approach could be to tell the user to feed via Heat parameter, at the time of the upgrade, their old default 1. https://github.com/openstack/tripleo-heat-templates/blob/master/puppet/hieradata /ceph.yaml
Bulk update to reflect scope of Red Hat OpenStack Platform 9 and Red Hat OpenStack Platform does not include this issue (No pm_ack+).
Instead of cloning for 11, I will reopen and target to OSP-11. Issue is lack of tracking of state for Ceph in Director... if we made the change in current architecture, on an upgrade this would be lost. Moving to 11, where the new architecture kicks in and hopefully we can address this.
default assignment as this is non-actionable. I don't see this as actionable for OPS 11 either.
Just looked into a fresh OSP10 deployment and this is what I'm getting: $ sudo df -h | grep osd Filesystem Size Used Avail Use% Mounted on /dev/vdb1 55G 38M 55G 1% /var/lib/ceph/osd/ceph-0 So it seems that the default osd location is honoured so moving this one to verified.
(In reply to seb from comment #11) > Just looked into a fresh OSP10 deployment and this is what I'm getting: > > > $ sudo df -h | grep osd > Filesystem Size Used Avail Use% Mounted on > /dev/vdb1 55G 38M 55G 1% /var/lib/ceph/osd/ceph-0 > > So it seems that the default osd location is honoured so moving this one to > verified. that is a symlink though and it's actually always been created, not sure if it is enough to close the BZ
@Giulio, no this is not a symlink, I also checked that on a fresh OSP10 install.
It is a symlink for me, unless you override the default OSDs config via hiera. # sudo hiera ceph::profile::params::osds {"/srv/data"=>{}} # ls -l /var/lib/ceph/osd/ total 0 lrwxrwxrwx. 1 root root 9 Oct 25 13:49 ceph-0 -> /srv/data
Hum I might have done then, alright.
OK, we still have it in the OSP 10 program for post-GA completion.