Hi,

Got a strange failure in our main VM environment using DP 23.4. We have a vCenter with over 2000 VM's, but are struggling to backup just two of them for some unknown reason. Please note we can't send logs to OpenText, so support on this forum is our only option.

Sequence of events:

• A couple of Windows DC VM's were rebuilt by our Windows support team to new Windows 2022. "Rebuild" was done by creating new VM's, named slightly differently from the original DC VM's which were deleted, and the new ones stored in the same datastores
• We tried to back them up but had problems with Changed Block Tracking DP settings with regard to a FULL or INCR backup
• Our default CBT setting it on for "Use changed block tracking", and "Allow fallback to non-CBT backups"
• But for some reason the VM's for a FULL would not work for CBT on, only when we disabled CBT would the FULL work. INCR would work with CBT on. Error would be a vepa_bar.exe error:
  • vepar_bar.exe: /svnstore/unix/panther/dp_git_master/bin/UNIX/boost/boost_1_65/include/boost/smart_ptr/shares_ptr.hpp:735: typename boost::detail::sp_member_access<T>::type boost::shared_ptr<T>::operator->() const [withT = VirtualEnvironment::ChangedBlockTracking; typename boost::detail::sp_member_access<T>::type = VirtualEnvironment::ChangedBlockTracking*]: Assertion 'px != o' failed.
  • None of the Disk Agents completed successfully
• So we decided to create two new empty datastores, and got each failed DC moved to their own datastore
• But now we have them in their own datastores, nothing works, FULL or INCR, whether CBT is on or off in the DP GUI! Error is different, it fails just as it gets to the stage in the job of trying to get changed blocks for CBT:
  • Virtual machine 'DC1': Could not backup disk scsci0:0
  • Backup of object failed
  • This is repeated for all three scsi disks
• When we run this job with debug we get these cryptic messages:
  • [110] [diskLibLog] VixDiskLib: VixDiskLib_OpenEx: Open a disk
  • [110] [diskLibLog] DISKLIB-LINK : "san://[specType:0,ssId:snapshot-13132342,spec:{vmxSpec:moref=vm-2424352}}[Datastore_name] DC1/DC1.vmdk@ip_addess:443?domain\userid/XXX : failed to open (Unknown error).
  • .......DISKLIB-CHAIN....DiskChainOpen.......failed to open: Unknown error
  • .......DISKLIB-LIB.....Unknown error (318)
  • ....VixDiskLib: Unable to locate appropriate transport mode to open disk. Error 13 (You do not have access rights to this file)......
• So this last message when the two VM's are put in to empty datastores is interesting, we use the same userid/password for all VM backups in this one vCenter, but for just these two, it's suggesting we do not have access rights, the VM's are set up the same as others, and permissions to our backup userid are the same as all the rest.

So we're a little lost now, we could try moving the VM's back to the shared datastores, so at least with some CBT changes in the DP GUI we could manually back them up. 

Any ideas anyone?

Cheers,

Andy