[ovirt-users] iSCSI Data Domain Down

[Clint Boggio]

Greetings oVirt Family;

Due to catastrophic power failure, my datacenter lost power. I am using a CentOS7 server to provide ISCSI services to my OVirt platform.

When the power came back on, and the iscsi server booted back up, the filters in lvm.conf were faulty and LVM assumed control over the LVM's that OVirt uses as the disks for the VMs. This tanked target.service because it claims "device already in use" and my datacenter is down.

I've tried several filter combinations in lvm.conf to no avail, and in my search I've found no documentation on how to make LVM "forget" about the volumes that it had assumed and release them.

Do any of you know of a procedure to make lvm forget about and release the volumes on /dev/sda ?

OVirt 3.6.5 on CentOS 7
4 Hypervisor nodes CentOS7
1 Dedicated engine CentOS7
1 iscsi SAN CentOS 7 exporting 10TB block device from a Dell Perc RAID controller /dev/sda with targetcli.
1 NFS server for ISO and Export Domains 5TB

I'm out I ideas and any help would be greatly appreciated. 

I'm currently using dd to recover the VM disk drives over to the NFS server in case this cannot be recovered.