Hi all,
I am having difficulty working out how OSDs are started automatically at
boot, as they are not currently in my simple deployment
Ubuntu 12.04
root@ceph-osd98:/usr/bin# ceph -v
ceph version 0.72.2 (a913ded2ff138aefb8cb84d347d72164099cfd60)
I gather this is the script I need to investigate:
/etc/init/ceph-osd-all-starter.conf
description "Ceph OSD (start all instances)"
start on starting ceph-osd-all
task
script
set -e
# first activate any partitions
ceph-disk activate-all
# TODO what's the valid charset for cluster names and osd ids?
find -L /var/lib/ceph/osd/ -mindepth 1 -maxdepth 1 -regextype posix-egrep
-regex '.*/[A-Za-z0-9]+-[A-Za-z0-9._-]+' -printf '%P\n' \
| while read f; do
if [ -e "/var/lib/ceph/osd/$f/ready" ] && [ -e
"/var/lib/ceph/osd/$f/upstart" ] && [ ! -e "/var/lib/ceph/osd/$f/sysvinit"
]; then
cluster="${f%%-*}"
id="${f#*-}"
initctl emit ceph-osd cluster="$cluster" id="$id"
fi
done
end script
Stepping through, I have this issue:
root@ceph-osd98:/usr/bin# ceph-disk activate-all
INFO:ceph-disk:Activating
/dev/disk/by-parttypeuuid/4fbd7e29-9d25-41b8-afd0-062c0ceff05d.55a9f1f8-8ef9-4420-8dc1-829ddff972e5
Error EINVAL: entity osd.0 exists but cap mon does not match
ERROR:ceph-disk:Failed to activate
ceph-disk: 'CalledProcessError' object has no attribute 'errno'
ceph-disk: Error: One or more partitions failed to activate
So it is finding the partition
root@ceph-osd98:/usr/bin# ceph auth get osd.0
exported keyring for osd.0
[osd.0]
key = AQDiQlhTuEF+KBAAUv4ClrlibNcFMv7wsR00Nw==
caps mon = "allow rwx"
caps osd = "allow *"
root@ceph-osd98:/usr/bin# less /var/lib/ceph/osd/ceph-0/keyring
[osd.0]
key = AQDiQlhTuEF+KBAAUv4ClrlibNcFMv7wsR00Nw==
caps were created using the manual deployment guide:
http://ceph.com/docs/master/install/manual-deployment/
sudo ceph auth add osd.{osd-num} osd 'allow *' mon 'allow rwx' -i
/var/lib/ceph/osd/ceph-{osd-num}/keyring
So I am not sure what is happening.
Where is this going wrong?
Cheers
Adam
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
