updating ceph-osd config

[James Page]

Hi Daniel

On Tue, 3 Jan 2017 at 03:58 Daniel Bidwell <drbidwell at gmail.com> wrote:

I started out installing openstack-base and then went back to add more
disks to be used for block storage.  The default was /dev/sdb.  I added
/dev/sda, /dev/sdc, and /dev/sdd via the juju gui, saved and deployed.
 Nothing happened.

I just went in and partitioned /dev/sda, /dev/sdc, and /dev/sdd on all
machines.  Using 2.0.2-xenial-amd64 how do I tell ceph-osd to go reread
the configuration and setup the additional disks?


The osd-devices configuration option acts as a whitelist of devices to use
for all ceph-osd units

Simply changing the configuration option should be sufficient - the
ceph-osd charm will attempt to detect new disks during the config-changed
hook.


There is a resolved/retry in the gui, but these don't have errors and
thus there is nothing to resolve/retry.


This would indicate that either the disks are in use, or that they where
not detected in some way.

It would be useful to see the unit logs for the ceph-osd units to determine
exactly what happened (/var/log/juju/unit-ceph-osd-*.log).

You should see information messages about which devices are detected and
whether they are in use or not.

Cheers

James
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/juju/attachments/20170103/73a8229e/attachment.html>