Initial Backup Fails of Large System
If an ABD backup of a large system fails, it may produce error messages such as "Detected stalled I/O, timeout reached (30 mins). Issuing cancellation request to ABD" or "Munge of disk [X] ended, but was still in progress".
This could be because the VM is large, and initial backup of large systems require special treatment. Because every environment's disk speed, network, and A2 host has different performance characteristics, there is no set definition for "large" VM. But generally, if your VM is over 4TB in size or has more than 4 disks, it may need special treatment during initial backup. There is no harm in attempting this procedure in any case.
The procedure is as follows:
- If necessary, create a new job for this VM that only included this VM, and no other VM.
- Select only ONE drive to protect for this VM. Start with the smallest disk first.
- Run the backup job. Wait for it to complete successfully.
- Select another disk to protect in the job, so the job has +1 disks selected. Continue in order of disk size, from smallest to largest.
- Repeat steps 3 and 4 until the entire VM is protected successfully.
If this approach does not work, please change the memory assigned to your ABD templates to 1GB of RAM. Then re-run the above steps. Typically, extra RAM is not necessary for your ABD. But there have been cases where this intervention appears to resolve the issue.