old VM resurrection weirdness

  • Thread starter Thread starter lonegroover
  • Start date Start date
L

lonegroover

Guest
Hi,

We run Proxmox VE 1.8 on two nodes arranged as a cluster with shared storage for the (KVM) VMs on two Cisco SANs. A few weeks ago one of the physical servers crashed. When it was restarted and its VMs brought back to life, something rather strange happened.

One of the VMs, a development server, appeared to lose all of its updates since August. Software that had been installed since then disappeared, along with any and all logfiles written since then.

The VM in question lives on drbdvg1-vm--114--disk--1, which is mirrored across the two SANs, of course.

Is it possible that one of the SANs has a more up-to-date version, and the cluster is seeing an invalid copy? I can't see any error messages in the logs to the effect that there is or was a synchronisation problem. Anyone seen anything like this? Or know how, if even possible, to recover from it?

Thanks for any help.