Hi, It looks like the node cannot find the vms' images (lvs) although it can see the domain, which is very strange. Is this an iSCSI or FCP domain? More logs would help understand exactly what is going on there. Also, could you please paste the output of: # lvs ce77262f-8346-42e5-823a-bd321f0814e7 on the problematic node and also from another node in the same cluster (if possible)? After getting the output of the lvs command, you can try to set the host to maintenance and then activate it again from oVirt Thanks, Gadi Ickowicz ----- Original Message ----- From: "Jakub Bittner" <j.bittner(a)nbu.cz&gt; To: users(a)ovirt.org Sent: Monday, February 10, 2014 6:55:40 PM Subject: [Users] node can not access disks Hi, after node restart oVirt can not start any virtual machine due to unknown storage error. Logs are here http://fpaste.org/75881/. It looks like it can not read LVM. Is there any way how to fix that? Thanks _______________________________________________ Users mailing list Users(a)ovirt.org http://lists.ovirt.org/mailman/listinfo/users

Hi, well at this stage it looks very strange: * it looks like the disks are really not on the storage domain - the lv for the vm you are attemtping to start (From the previous mail's logs) is not visible at all. * did something happen to the storage server itself/is it possible that things were deleted from the storage itself? (maybe you deleted the vms' disks?) * On the problematic node I see that the lvm command gave some errors regarding missing devices: /dev/mapper/360a9800042415569305d434565795a54: read failed after 0 of 4096 at 10737352704: Chyba vstupu/výstupu /dev/mapper/360a9800042415569305d434565795a54: read failed after 0 of 4096 at 10737410048: Chyba vstupu/výstupu /dev/mapper/360a9800042415569305d434565795a54: read failed after 0 of 4096 at 0: Chyba vstupu/výstupu /dev/mapper/360a9800042415569305d434565795a54: read failed after 0 of 4096 at 4096: Chyba vstupu/výstupu * Is the domain a single LUN or multiple LUNs? what is the output of # multipath -ll # vgs could you attach the full vdsm log from the problematic host maybe? Gadi Ickowicz ----- Original Message ----- From: "Jakub Bittner" <j.bittner(a)nbu.cz&gt; To: "Gadi Ickowicz" <gickowic(a)redhat.com&gt; Cc: users(a)ovirt.org Sent: Tuesday, February 11, 2014 10:25:33 AM Subject: Re: [Users] node can not access disks Dne 11.2.2014 07:47, Gadi Ickowicz napsal(a): Hello, it is ISCSI domain. All nodes are problematic. lvs output: http://fpaste.org/76058/ I restarted problematic node but it did not help. So I removed it from cluster, but many other VMs on ther nodes are dead (can not run, or can not boot from disk because they do not see it). About 5% of VM we were able to export or restart to stable status