I'm running a couple of CentOS4.5 boxes with all current updates configured as a cluster. Both have access to the same three iscsi-connected disks on which I've created a mirrored LVM volume. I've created a GFS filesystem on the LVM volume and can mount it and read/write from both machines in the cluster. So far, so good. Now I want to test what happens when different parts of the system fail. First I thought I'd try disconnecting one of the iscsi disks that form the data component of the LVM mirror, leaving the other data disk and log active. I expected LVM to fail the disk and run in degraded linear mode until I restored the disconnected disk. What actually happens is that all access to the LVM volume hangs. The iscsi layer on the cluster servers logs that the connection has dropped, but that's it. Am I missing something important here, or is this just expecting too much? James -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
