Hi Florent, On 22/12/2014 19:49, Florent MONTHEL wrote: > Hi Loic, Hi Robert, > > Thanks. I’m integrating CEPH OSD with OpenSVC services (http://www.opensvc.com) so I have to generate UUID myself in order to map services > It’s the reason for that I’m generating sgdisk commands with my own UUID > > After activating OSD, I don’t have mapping osd & journal with cep-disk command > > root@raven:/var/lib/ceph/osd/ceph-5# ceph-disk list > /dev/sda other, ext4, mounted on / > /dev/sdb swap, swap > /dev/sdc : > /dev/sdc1 ceph journal > /dev/sdd : > /dev/sdd1 ceph data, active, cluster ceph, osd.3 > /dev/sde : > /dev/sde1 ceph journal > /dev/sdf : > /dev/sdf1 ceph data, active, cluster ceph, osd.4 > /dev/sdg : > /dev/sdg1 ceph journal > /dev/sdh : > /dev/sdh1 ceph data, active, cluster ceph, osd.5 > > After below command (osd 5), ceph-deploy didn’t create file "journal_uuid" : > > ceph-deploy --overwrite-conf osd create raven:/dev/disk/by-partuuid/6356fd8d-0d84-432a-b9f4-3d02f94afdff:/dev/disk/by-partuuid/36741e5b-eee0-4368-9736-a31701a186a1 > > root@raven:/var/lib/ceph/osd/ceph-5# ls -l > total 56 > -rw-r--r-- 1 root root 192 Dec 21 23:55 activate.monmap > -rw-r--r-- 1 root root 3 Dec 21 23:55 active > -rw-r--r-- 1 root root 37 Dec 21 23:55 ceph_fsid > drwxr-xr-x 184 root root 8192 Dec 22 19:25 current > -rw-r--r-- 1 root root 37 Dec 21 23:55 fsid > lrwxrwxrwx 1 root root 58 Dec 21 23:55 journal -> /dev/disk/by-partuuid/36741e5b-eee0-4368-9736-a31701a186a1 > -rw------- 1 root root 56 Dec 21 23:55 keyring > -rw-r--r-- 1 root root 21 Dec 21 23:55 magic > -rw-r--r-- 1 root root 6 Dec 21 23:55 ready > -rw-r--r-- 1 root root 4 Dec 21 23:55 store_version > -rw-r--r-- 1 root root 53 Dec 21 23:55 superblock > -rw-r--r-- 1 root root 0 Dec 22 19:24 sysvinit > -rw-r--r-- 1 root root 2 Dec 21 23:55 whoami > > > So I created for each osd, file "journal_uuid » manually and mapping become OK with ceph-disk :) > > root@raven:/var/lib/ceph/osd/ceph-5# echo "36741e5b-eee0-4368-9736-a31701a186a1 » > journal_uuid I think this is an indication that when you ceph-disk prepare the device the journal_uuid was not provided and therefore the journal_uuid creation was skipped: http://workbench.dachary.org/ceph/ceph/blob/giant/src/ceph-disk#L1235 called from http://workbench.dachary.org/ceph/ceph/blob/giant/src/ceph-disk#L1338 Cheers > > It’s ok now : > > root@raven:/var/lib/ceph/osd/ceph-5# ceph-disk list > /dev/sda other, ext4, mounted on / > /dev/sdb swap, swap > /dev/sdc : > /dev/sdc1 ceph journal, for /dev/sdd1 > /dev/sdd : > /dev/sdd1 ceph data, active, cluster ceph, osd.3, journal /dev/sdc1 > /dev/sde : > /dev/sde1 ceph journal, for /dev/sdf1 > /dev/sdf : > /dev/sdf1 ceph data, active, cluster ceph, osd.4, journal /dev/sde1 > /dev/sdg : > /dev/sdg1 ceph journal, for /dev/sdh1 > /dev/sdh : > /dev/sdh1 ceph data, active, cluster ceph, osd.5, journal /dev/sdg1 > > > Thanks robert@xxxxxxxxxxxxx <mailto:robert@xxxxxxxxxxxxx> for clue ;) > > *Florent Monthel** > * > > > > > >> Le 21 déc. 2014 à 18:08, Loic Dachary <loic@xxxxxxxxxxx <mailto:loic@xxxxxxxxxxx>> a écrit : >> >> Hi Florent, >> >> It is unusual to manually run the sgdisk. Is there a reason why you need to do this instead of letting ceph-disk prepare do it for you ? >> >> The information about the association between journal and data is only displayed when the OSD has been activated. See http://workbench.dachary.org/ceph/ceph/blob/giant/src/ceph-disk#L2246 >> Cheers >> >> On 21/12/2014 15:11, Florent MONTHEL wrote: >>> Hi, >>> >>> I would like to separate OSD and journal on 2 différent disks so I have : >>> >>> 1 disk /dev/sde (1GB) for journal => type code JOURNAL_UUID = '45b0969e-9b03-4f30-b4c6-b4b80ceff106' >>> 1 disk /dev/sdd (5GB) for OSD => type code OSD_UUID = '4fbd7e29-9d25-41b8-afd0-062c0ceff05d' >>> >>> I execute below commands : >>> >>> FOR JOURNAL : >>> sgdisk --new=1:0:1023M --change-name="1:ceph journal" --partition-guid=1:e89f18cc-ae46-4573-8bca-3e782d45849c --typecode=1:45b0969e-9b03-4f30-b4c6-b4b80ceff106 -- /dev/sde >>> >>> FOR OSD: >>> sgdisk --new=1:0:5119M --change-name="1:ceph data" --partition-guid=1:7476f0a8-a6cd-4224-b64b-a4834c32a73e --typecode=1:4fbd7e29-9d25-41b8-afd0-062c0ceff05d -- /dev/sdd >>> >>> And I'm preparing OSD : >>> ceph-disk prepare --osd-uuid 7476f0a8-a6cd-4224-b64b-a4834c32a73e --journal-uuid e89f18cc-ae46-4573-8bca-3e782d45849c --fs-type xfs --cluster ceph -- /dev/sdd1 /dev/sde1 >>> >>> >>> After that, I don’t see relation between /dev/sde1 & /dev/sdc1 >>> >>> root@falcon:/srv/ceph01adm001/data/cluster-ceph01# ceph-disk list >>> /dev/sdd : >>> /dev/sdd1 ceph data, prepared, cluster ceph >>> /dev/sde : >>> /dev/sde1 ceph journal >>> >>> Is it normal ? >>> >>> Thanks >>> >>> >>> *Florent Monthel** >>> * >>> >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> ceph-users mailing list >>> ceph-users@xxxxxxxxxxxxxx <mailto:ceph-users@xxxxxxxxxxxxxx> >>> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com >>> >> >> -- >> Loïc Dachary, Artisan Logiciel Libre >> > -- Loïc Dachary, Artisan Logiciel Libre
Attachment:
signature.asc
Description: OpenPGP digital signature
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
