Thanks for the hint guys :) Running the following commands on all nodes resolved the issue: chmod u+rxw /data/brick*/brick* chmod g+rx /data/brick*/brick* I wonder if the upgrade process from gluster.org to centos.org packages caused this issue. I have another oVirt cluster running in a data center and it has similar package versions before this upgrade. I can try to upgrade one of the nodes there next week and see if the directory permission issue happens again. I will notify the gluster list and/or open a bug report for CentOS package maintainers when I finish my test. On Thu, Jul 28, 2016 at 9:29 PM David Gossage <dgossage@carouselchecks.com> wrote:
On Thu, Jul 28, 2016 at 11:44 AM, Siavash Safi <siavash.safi@gmail.com> wrote:
It seems that dir modes are wrong!? [root@node1 ~]# ls -ld /data/brick*/brick* drw-------. 5 vdsm kvm 107 Jul 28 20:13 /data/brick1/brick1 drw-------. 5 vdsm kvm 82 Jul 27 23:08 /data/brick2/brick2 [root@node2 ~]# ls -ld /data/brick*/brick* drwxr-xr-x. 5 vdsm kvm 107 Apr 26 19:33 /data/brick1/brick1 drw-------. 5 vdsm kvm 82 Jul 27 23:08 /data/brick2/brick2 drw-------. 5 vdsm kvm 107 Jul 28 20:13 /data/brick3/brick3 [root@node3 ~]# ls -ld /data/brick*/brick* drw-------. 5 vdsm kvm 107 Jul 28 20:10 /data/brick1/brick1 drw-------. 5 vdsm kvm 82 Jul 27 23:08 /data/brick2/brick2
That would probably do it. kvm does read access. plus the lack of x on directories isnt great either. I'd think since they are the bricks you could maybe manually chmod them appropriately 755. I "think" gluster is only tracking files under /data/brick1/brick1/* not /data/brick1/brick1 itself.
/data/brick3/brick3 is the only "new" one? Wonder if it could be a bug from the brick move in some way. Might be something worth posting to gluster list about.
On Thu, Jul 28, 2016 at 9:06 PM Sahina Bose <sabose@redhat.com> wrote:
----- Original Message -----
From: "Siavash Safi" <siavash.safi@gmail.com> To: "Sahina Bose" <sabose@redhat.com> Cc: "David Gossage" <dgossage@carouselchecks.com>, "users" < users@ovirt.org>, "Nir Soffer" <nsoffer@redhat.com>, "Allon Mureinik" <amureini@redhat.com> Sent: Thursday, July 28, 2016 9:04:32 PM Subject: Re: [ovirt-users] Cannot find master domain
Please check the attachment.
Nothing out of place in the mount logs.
Can you ensure the brick dir permissions are vdsm:kvm - even for the brick that was replaced?
On Thu, Jul 28, 2016 at 7:46 PM Sahina Bose <sabose@redhat.com> wrote:
----- Original Message -----
From: "Siavash Safi" <siavash.safi@gmail.com> To: "Sahina Bose" <sabose@redhat.com> Cc: "David Gossage" <dgossage@carouselchecks.com>, "users" < users@ovirt.org> Sent: Thursday, July 28, 2016 8:35:18 PM Subject: Re: [ovirt-users] Cannot find master domain
[root@node1 ~]# ls -ld /rhev/data-center/mnt/glusterSD/ drwxr-xr-x. 2 vdsm kvm 6 Jul 28 19:28
[root@node1 ~]# getfacl /rhev/data-center/mnt/glusterSD/ getfacl: Removing leading '/' from absolute path names # file: rhev/data-center/mnt/glusterSD/ # owner: vdsm # group: kvm user::rwx group::r-x other::r-x
The ACLs look correct to me. Adding Nir/Allon for insights.
Can you attach the gluster mount logs from this host?
And as I mentioned in another message, the directory is empty.
On Thu, Jul 28, 2016 at 7:24 PM Sahina Bose <sabose@redhat.com> wrote:
> Error from vdsm log: Permission settings on the specified path do not > allow access to the storage. Verify permission settings on the specified > storage path.: 'path = /rhev/data-center/mnt/glusterSD/ 172.16.0.11: _ovirt' > > I remember another thread about a similar issue - can you check
> settings on the storage path? > > ----- Original Message ----- > > From: "Siavash Safi" <siavash.safi@gmail.com> > > To: "David Gossage" <dgossage@carouselchecks.com> > > Cc: "users" <users@ovirt.org> > > Sent: Thursday, July 28, 2016 7:58:29 PM > > Subject: Re: [ovirt-users] Cannot find master domain > > > > > > > > On Thu, Jul 28, 2016 at 6:29 PM David Gossage < > dgossage@carouselchecks.com > > > wrote: > > > > > > > > On Thu, Jul 28, 2016 at 8:52 AM, Siavash Safi < siavash.safi@gmail.com > > > wrote: > > > > > > > > Hi, > > > > Issue: Cannot find master domain > > Changes applied before issue started to happen: replaced > > 172.16.0.12:/data/brick1/brick1 with 172.16.0.12: /data/brick3/brick3, > did > > minor package upgrades for vdsm and glusterfs > > > > vdsm log: https://paste.fedoraproject.org/396842/ > > > > > > Any errrors in glusters brick or server logs? The client gluster logs > from > > ovirt? > > Brick errors: > > [2016-07-28 14:03:25.002396] E [MSGID: 113091] [posix.c:178:posix_lookup] > > 0-ovirt-posix: null gfid for path (null) > > [2016-07-28 14:03:25.002430] E [MSGID: 113018] [posix.c:196:posix_lookup] > > 0-ovirt-posix: lstat on null failed [Invalid argument] > > (Both repeated many times) > > > > Server errors: > > None > > > > Client errors: > > None > > > > > > > > > > > > > > > > yum log: https://paste.fedoraproject.org/396854/ > > > > What version of gluster was running prior to update to 3.7.13? > > 3.7.11-1 from gluster.org repository(after update ovirt switched to > centos > > repository) > > > > > > > > > > Did it create gluster mounts on server when attempting to start? > > As I checked the master domain is not mounted on any nodes. > > Restarting vdsmd generated following errors: > > > > jsonrpc.Executor/5::DEBUG::2016-07-28 > > 18:50:57,661::fileUtils::143::Storage.fileUtils::(createdir) Creating > > directory: /rhev/data-center/mnt/glusterSD/172.16.0.11:_ovirt mode: None > > jsonrpc.Executor/5::DEBUG::2016-07-28 > > >
18:50:57,661::storageServer::364::Storage.StorageServer.MountConnection::(_get_backup_servers_option)
> > Using bricks: ['172.16.0.11', '172.16.0.12', '172.16.0.13'] > > jsonrpc.Executor/5::DEBUG::2016-07-28 > > 18:50:57,662::mount::229::Storage.Misc.excCmd::(_runcmd) /usr/bin/taskset > > --cpu-list 0-31 /usr/bin/sudo -n /usr/bin/systemd-run --scope > > --slice=vdsm-glusterfs /usr/bin/mount -t glusterfs -o > > backup-volfile-servers=172.16.0.12:172.16.0.13 172.16.0.11: /ovirt > > /rhev/data-center/mnt/glusterSD/172.16.0.11:_ovirt (cwd None) > > jsonrpc.Executor/5::DEBUG::2016-07-28 > > 18:50:57,789::__init__::318::IOProcessClient::(_run) Starting > IOProcess... > > jsonrpc.Executor/5::DEBUG::2016-07-28 > > 18:50:57,802::mount::229::Storage.Misc.excCmd::(_runcmd) /usr/bin/taskset > > --cpu-list 0-31 /usr/bin/sudo -n /usr/bin/umount -f -l > > /rhev/data-center/mnt/glusterSD/172.16.0.11:_ovirt (cwd None) > > jsonrpc.Executor/5::ERROR::2016-07-28 > > 18:50:57,813::hsm::2473::Storage.HSM::(connectStorageServer) Could not > > connect to storageServer > > Traceback (most recent call last): > > File "/usr/share/vdsm/storage/hsm.py", line 2470, in connectStorageServer > > conObj.connect() > > File "/usr/share/vdsm/storage/storageServer.py", line 248, in connect > > six.reraise(t, v, tb) > > File "/usr/share/vdsm/storage/storageServer.py", line 241, in connect > > self.getMountObj().getRecord().fs_file) > > File "/usr/share/vdsm/storage/fileSD.py", line 79, in validateDirAccess > > raise se.StorageServerAccessPermissionError(dirPath) > > StorageServerAccessPermissionError: Permission settings on the specified > path > > do not allow access to the storage. Verify permission settings on the > > specified storage path.: 'path = > > /rhev/data-center/mnt/glusterSD/172.16.0.11:_ovirt' > > jsonrpc.Executor/5::DEBUG::2016-07-28 > > 18:50:57,817::hsm::2497::Storage.HSM::(connectStorageServer) knownSDs: {} > > jsonrpc.Executor/5::INFO::2016-07-28 > > 18:50:57,817::logUtils::51::dispatcher::(wrapper) Run and
/rhev/data-center/mnt/glusterSD/ the ACL protect:
> > connectStorageServer, Return response: {'statuslist': [{'status': 469, > 'id': > > u'2d285de3-eede-42aa-b7d6-7b8c6e0667bc'}]} > > jsonrpc.Executor/5::DEBUG::2016-07-28 > > 18:50:57,817::task::1191::Storage.TaskManager.Task::(prepare) > > Task=`21487eb4-de9b-47a3-aa37-7dce06533cc9`::finished: {'statuslist': > > [{'status': 469, 'id': u'2d285de3-eede-42aa-b7d6-7b8c6e0667bc'}]} > > jsonrpc.Executor/5::DEBUG::2016-07-28 > > 18:50:57,817::task::595::Storage.TaskManager.Task::(_updateState) > > Task=`21487eb4-de9b-47a3-aa37-7dce06533cc9`::moving from state preparing > -> > > state finished > > > > I can manually mount the gluster volume on the same server. > > > > > > > > > > > > > > > > > > > > > > > > > > > > Setup: > > engine running on a separate node > > 3 x kvm/glusterd nodes > > > > Status of volume: ovirt > > Gluster process TCP Port RDMA Port Online Pid > > >
> > Brick 172.16.0.11:/data/brick1/brick1 49152 0 Y 17304 > > Brick 172.16.0.12:/data/brick3/brick3 49155 0 Y 9363 > > Brick 172.16.0.13:/data/brick1/brick1 49152 0 Y 23684 > > Brick 172.16.0.11:/data/brick2/brick2 49153 0 Y 17323 > > Brick 172.16.0.12:/data/brick2/brick2 49153 0 Y 9382 > > Brick 172.16.0.13:/data/brick2/brick2 49153 0 Y 23703 > > NFS Server on localhost 2049 0 Y 30508 > > Self-heal Daemon on localhost N/A N/A Y 30521 > > NFS Server on 172.16.0.11 2049 0 Y 24999 > > Self-heal Daemon on 172.16.0.11 N/A N/A Y 25016 > > NFS Server on 172.16.0.13 2049 0 Y 25379 > > Self-heal Daemon on 172.16.0.13 N/A N/A Y 25509 > > > > Task Status of Volume ovirt > > >
> > Task : Rebalance > > ID : 84d5ab2a-275e-421d-842b-928a9326c19a > > Status : completed > > > > Thanks, > > Siavash > > > > > > > > _______________________________________________ > > Users mailing list > > Users@ovirt.org > > http://lists.ovirt.org/mailman/listinfo/users > > > > > > > > _______________________________________________ > > Users mailing list > > Users@ovirt.org > > http://lists.ovirt.org/mailman/listinfo/users > > >