# CEPH and rbd usage

We are using CEPH to provide some rbd volumes to our ganeti cluster.

First, it's useful to have a list of all rbd volumes and snapshots in cluster:

root@r1u28:/srv/gnt-info# ./gnt-rbd.sh
docker1 8bd20a37-b1f8-4c58-aa45-1f3cb8bb859a.rbd.disk0
gray 4ee9a9ee-5081-4965-ba5e-ed2a8cf9d97c.rbd.disk0
mudrac 78ad4b47-fd52-44b1-ae9b-b8fffc8c1cc5.rbd.disk0
mudrac 214e17fa-22e4-490d-9ef7-d07d666f6ddf.rbd.disk1
mudrac 0f4e2fb9-09e5-4459-a381-1f07ba5a98f5.rbd.disk2
mudrac f734a9d5-0c81-4840-a798-951941f597da.rbd.disk3
odin.ffzg.hr 31fa150f-8ab3-49bf-8924-7b2b1caece52.rbd.disk0
odin.ffzg.hr 31fa150f-8ab3-49bf-8924-7b2b1caece52.rbd.disk0@2017-07-20 #30 3072 MB
odin.ffzg.hr eec7e8bf-809e-4d69-a929-49b6fee3227f.rbd.disk1
odin.ffzg.hr eec7e8bf-809e-4d69-a929-49b6fee3227f.rbd.disk1@2017-07-20 #31 250 GB
syslog 45b765f6-316c-46a1-a92a-071113e18fcb.rbd.disk0
video 7cda803e-f877-4001-b9ce-9d460e3f85b8.rbd.disk0
video f1991f16-f59c-4e9d-adae-79cbfe034878.rbd.disk1

Then there is script to start snapshot backup from rbd instance to lib15:

root@r1u28:/srv/gnt-info# ./rbd-snap-backup.sh odin.ffzg.hr 1

gnt-lv-snap-rsync.sh will call this script if it can't find any lvm volumes
for that disk

If backup fails for some reason, it might leave mounted filesystem. You have
to umount that manually. After that, there is cleanup script to remove left-overs
of @backup snapshots from ceph:

root@r1u28:/srv/gnt-info# ./rbd-remove-snap.sh