Hi folks, I've been using proxmox to great satisfaction this last year as my main hypervisor, but I've been having a very weird specific issue with an old Oracle 11g Database inside an Ubuntu 12.04 VM that's part of a legacy system that needs to keep running (It's always those, isn't it?). The VM is a bone stock default VM, default CPU, i440fx machine type, ZFS storage.
Basically, it was hosted on an older PVE 6.x (the final packages for 6) server we're trying to decommission and repurpose, and it all ran fine there for years; however, once migrating to the new shiny new VM host running PVE7, with the same exact configuration - it was copied from the old server, along with the ZFS pool - I'm having the database just crash on me periodically with ORA-600 errors. Luckily the old server hasn't been formatted yet, so I gave it a couple of days to run from there with no issues.
Today, I backupped on the old hypervisor and restored on the new one, thinking maybe it was a problem with the ZFS transfer, but no dice. Instance keeps shutting down periodically.
Last thing I tried was changing machine type to pc-i440fx-6.2 (trying to mimic the older QEMU version in PVE 6.2) but no dice, it's crashed once already. I'm about to try specifying a CPU type for it (possibly Haswell, where it was hosted before?) but other than that I'm out of ideas. Has anyone ever dealt with something like this?
Basically, it was hosted on an older PVE 6.x (the final packages for 6) server we're trying to decommission and repurpose, and it all ran fine there for years; however, once migrating to the new shiny new VM host running PVE7, with the same exact configuration - it was copied from the old server, along with the ZFS pool - I'm having the database just crash on me periodically with ORA-600 errors. Luckily the old server hasn't been formatted yet, so I gave it a couple of days to run from there with no issues.
Today, I backupped on the old hypervisor and restored on the new one, thinking maybe it was a problem with the ZFS transfer, but no dice. Instance keeps shutting down periodically.
Last thing I tried was changing machine type to pc-i440fx-6.2 (trying to mimic the older QEMU version in PVE 6.2) but no dice, it's crashed once already. I'm about to try specifying a CPU type for it (possibly Haswell, where it was hosted before?) but other than that I'm out of ideas. Has anyone ever dealt with something like this?
