vRDM and CBT
I have a vRDM that is approximately 1.5 TB. After reading some documentation and talking to a few DP representatives vRDMs are supported by DP, but after doing a full backup and a subsequent incremental, it has made me question my choice of using a vRDM (originally my manager had elected to go with a vRDM to avoid overhead by the VMFS).
I see a cbk file on the data store for the corresponding vRDM, however it appears that my subsequent vRDM incremental backups are the same size as the full. Is this an expected behavior (rule out the behavior of any applications on the vRDM; assume static content).
So here's a status update on this post. I've engaged both VMware and HP for troubleshooting help. In VMware's opinion, CBT is enabled on vRDMs and therefore it should be possible to do incremental backups on volumes that are mapped directly to VMs in vRDM mode. Per some digging around, HP seems to "think" they are gathering the CBT blocks correctly but VMware's VDDK is not giving them the right information. Upon reviewing HP's notes VMware has forwarded the specific case to the VDDK team and they are now waiting to hear from HP's DP Lab on their method of access and how they interpret CBTs provided by the VDDK. If anyone from HP or VMware lurks this forum, I can provide SRs for VMware and ref#s from HP for the information that's already been assembled for this issue.
So here's an update. Thanks to whoever unlocked this post!
- DP has acknowledged that there is a problem with the way they are handling the Change Blocks even though the problem is seemingly VMware's issue.
- DP Lab group has proved they do have the ability to do an incremental backup of vRDMs however, there are still a lot of kinks to work out.
- There are a total of 7 files they are messing with. They are located in ...\Program Files\Omniback\bin\. They are DPComServer.exe, vCloudAPI.dll, vepa_bar.exe, vepa_util.exe, vepalib_hyperv.dll, vepalib_vcd.dll, and vepalib_vmware.dll
The new Binaries allow me to do incremental backups of vRDM drives, however there are issues that I am currently addressing:
- Full, veagent backups are currently giving the Lost connection to OB2BAR Backup DA named "ERROR" on host insert server name here. Ipc subsystem reports: "IPC Read Error System Error:  Connection reset by peer". This seems to be directly related to the new binaries as the old 7.01 binaries do not exhibit this issue. Currently I am jockeying between binaries when I need to do backups
- Backup sessions fail to terminate. Backups get completely done, however the session never ends until the server restarts. This leaves the SAN lock reservation in the c:\windows\temp directory so unless you remember to delete the created SAN lock folder your backups will go over NBD until you physically remove the SAN lock.
- I have not attempted to do a restore yet.