Re: [Users] Testing gluster in 3.2 nightlies
On 4-1-2013 7:03, Kanagaraj wrote: On Friday 04 January 2013 02:12 AM, Joop wrote: Just saw the question what route to follow for post 3.2 and picked up something that I didn't know about was going to ask if it was possible to implement, namely setting permissions on the volume folder when creating a gluster volume. But when trying it out I found a little snag. I installed two hosts with a minimal fed17 install, added a couple of repos to enable installing the latest nightlies (it needed a later then 0.9 libvirt (0.10 from danken) and qemu-kvm and since a couple of days a later gluster which outputs in xml (gluster-3.4..) Steps: - start with a clean engine-setup - in the webui, set the cluster to 3.2 and virt+gluster, set the DC to 3.2/NFS - add the two hosts, went OK. - add a Volume, gluster-data, gets created OK, BUT the Create Volume 'popup window' never stops spinning its busy cursor :-( Looking at the logs I see that the volume is created, st01 is used, st02 gets the message, creation is returning Succes but it looks like the global lock is never released. We have an issue with locking while creating a volume, fix for this is currently under review. http://gerrit.ovirt.org/10448 I still might have the vdsm logs or else I can replay it to generate them, but if it really is this issue then it they might not be needed. Is the workaround I used a viable one? (restarting ovirt-engine) Joop ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Testing gluster in 3.2 nightlies
On Friday 04 January 2013 02:15 PM, noc wrote: On 4-1-2013 7:03, Kanagaraj wrote: On Friday 04 January 2013 02:12 AM, Joop wrote: Just saw the question what route to follow for post 3.2 and picked up something that I didn't know about was going to ask if it was possible to implement, namely setting permissions on the volume folder when creating a gluster volume. But when trying it out I found a little snag. I installed two hosts with a minimal fed17 install, added a couple of repos to enable installing the latest nightlies (it needed a later then 0.9 libvirt (0.10 from danken) and qemu-kvm and since a couple of days a later gluster which outputs in xml (gluster-3.4..) Steps: - start with a clean engine-setup - in the webui, set the cluster to 3.2 and virt+gluster, set the DC to 3.2/NFS - add the two hosts, went OK. - add a Volume, gluster-data, gets created OK, BUT the Create Volume 'popup window' never stops spinning its busy cursor :-( Looking at the logs I see that the volume is created, st01 is used, st02 gets the message, creation is returning Succes but it looks like the global lock is never released. We have an issue with locking while creating a volume, fix for this is currently under review. http://gerrit.ovirt.org/10448 I still might have the vdsm logs or else I can replay it to generate them, but if it really is this issue then it they might not be needed. Is the workaround I used a viable one? (restarting ovirt-engine) We are waiting for the fix to be approved. I can't think of any workaround for this now. After restarting also you will have the same issue while creating the next volume in the cluster. Other operations like start/stop/add-bricks will work after the restart. Joop ___ 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
Re: [Users] Testing gluster in 3.2 nightlies
On Friday 04 January 2013 02:12 AM, Joop wrote: Just saw the question what route to follow for post 3.2 and picked up something that I didn't know about was going to ask if it was possible to implement, namely setting permissions on the volume folder when creating a gluster volume. But when trying it out I found a little snag. I installed two hosts with a minimal fed17 install, added a couple of repos to enable installing the latest nightlies (it needed a later then 0.9 libvirt (0.10 from danken) and qemu-kvm and since a couple of days a later gluster which outputs in xml (gluster-3.4..) Steps: - start with a clean engine-setup - in the webui, set the cluster to 3.2 and virt+gluster, set the DC to 3.2/NFS - add the two hosts, went OK. - add a Volume, gluster-data, gets created OK, BUT the Create Volume 'popup window' never stops spinning its busy cursor :-( Looking at the logs I see that the volume is created, st01 is used, st02 gets the message, creation is returning Succes but it looks like the global lock is never released. We have an issue with locking while creating a volume, fix for this is currently under review. http://gerrit.ovirt.org/10448 At 20:09 I haven't done anything yet and then there is at 20:10 the create volume: [2013-01-03 20:09:58.937041] I [glusterd-handler.c:912:glusterd_handle_cli_get_volume] 0-glusterd: Received get vol req [2013-01-03 20:09:58.937949] W [socket.c:501:__socket_rwv] 0-socket.management: readv failed (No data available) [2013-01-03 20:10:03.968113] I [glusterd-handler.c:912:glusterd_handle_cli_get_volume] 0-glusterd: Received get vol req [2013-01-03 20:10:03.968991] W [socket.c:501:__socket_rwv] 0-socket.management: readv failed (No data available) [2013-01-03 20:10:08.733761] I [glusterd-volume-ops.c:76:glusterd_handle_create_volume] 0-management: Received create volume req [2013-01-03 20:10:08.739702] I [glusterd-utils.c:275:glusterd_lock] 0-glusterd: Cluster lock held by 9cbdd346-9296-4f98-bb69-ae226547abea [2013-01-03 20:10:08.739710] I [glusterd-handler.c:474:glusterd_op_txn_begin] 0-management: Acquired local lock [2013-01-03 20:10:08.740435] I [glusterd-rpc-ops.c:538:glusterd_cluster_lock_cbk] 0-glusterd: Received ACC from uuid: 367dd996-427a-45c4-bc23-73ac3f238367 [2013-01-03 20:10:08.764768] I [glusterd-op-sm.c:2583:glusterd_op_ac_send_stage_op] 0-glusterd: Sent op req to 1 peers [2013-01-03 20:10:08.792813] I [glusterd-rpc-ops.c:682:glusterd_stage_op_cbk] 0-glusterd: Received ACC from uuid: 367dd996-427a-45c4-bc23-73ac3f238367 [2013-01-03 20:10:09.802373] I [glusterd-utils.c:884:glusterd_volume_b The last line isn't an accident that is how the log ends after more then 15min of waiting. To get any further I can press ESC in the create volume window and then I need to restart ovirt-engine and can proceed. [root@st01 ~]# rpm -aq | grep gluster glusterfs-fuse-3.4.0qa6-1.el6.x86_64 vdsm-gluster-4.10.3-0.50.gitc6625ce.fc17.noarch glusterfs-3.4.0qa6-1.el6.x86_64 glusterfs-server-3.4.0qa6-1.el6.x86_64 [root@st01 ~]# rpm -aq | grep vdsm vdsm-4.10.3-0.50.gitc6625ce.fc17.x86_64 vdsm-xmlrpc-4.10.3-0.50.gitc6625ce.fc17.noarch vdsm-cli-4.10.3-0.50.gitc6625ce.fc17.noarch vdsm-gluster-4.10.3-0.50.gitc6625ce.fc17.noarch vdsm-python-4.10.3-0.50.gitc6625ce.fc17.x86_64 [root@st01 ~]# rpm -aq | grep libvirt libvirt-daemon-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-interface-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-nwfilter-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-xen-0.10.1-2.fc17.x86_64 libvirt-client-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-network-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-qemu-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-secret-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-nodedev-0.10.1-2.fc17.x86_64 libvirt-daemon-config-network-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-storage-0.10.1-2.fc17.x86_64 libvirt-lock-sanlock-0.10.1-2.fc17.x86_64 libvirt-0.10.1-2.fc17.x86_64 libvirt-python-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-lxc-0.10.1-2.fc17.x86_64 libvirt-daemon-config-nwfilter-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-uml-0.10.1-2.fc17.x86_64 libvirt-daemon-driver-libxl-0.10.1-2.fc17.x86_64 ___ 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