CaseID 00452680 Backup Copy WAN ACC Invalid length of binary

[VladV]

I recently opened a case for the following problem:

When using Backup Copy with wan acceleration i get the following error for a specific VM with a specific vmdk:

"Invalid length of the binary data '56025771'. Failed to extract binary content from buffer. Sync task 'e61a66d9-da04-4c82-8611-618295180134' initialization failed."

The VM has two disks, 64GB and 2TB. They are standard MBR disks, which backup and replicate without any issues.

Setup:
- Source:
VM 8 x vCPU / 18GB RAM
Windows Server 2012
Repository iSCSI 7TB Drive

- Target:
Physical 2 x 4 core CPU / 12GB RAM
Windows 7 x64 SP1
Repository SATA 4TB drive

- WAN 50 Mbps VPN link

The job failes processing this VM after ~ 4 retries, I believe, and then successfully processes the rest. If I disable WAN acceleration then it manages to process this VM too.

I tried many things, a lot I lost track of, but they included:

- reinstalling the WAN accelerator on both source and target
- reinstalling the backup repository on the target
- formatted the target disk
- added the VMs from infrastructure and backups
- clearing the cache on both source and target
- reduced the upload streams to 3

It's hard to get a sense of what's happening because the error message doesn't tell you much but I hope the support team will help me solve this problem soon.

Regards,
Vlad

[VladV]

Update: The R&D department confirmed this issue and hopefully this will be fixed in patch #2. Unfortunately, in my case, this makes the investment in the Enterprise Plus version a bit useless because we cannot use this functionality on one of our most important and largest VMs.

Regards,
Vlad

[foggy]

Vlad, thanks for coming back to update the thread and also for reporting the issue and working with support on its resolution, very much appreciated!

[veremin]

- reinstalling the WAN accelerator on both source and target
- reinstalling the backup repository on the target
- formatted the target disk
- added the VMs from infrastructure and backups
- clearing the cache on both source and target
- reduced the upload streams to 3

Just out of curiosity – have you also tried to map backup copy job or seed the global cache? Thanks.

[VladV]

I did this in the beginning.

As a history I first used the backup copy on a standalone eSATA 4TB disk on a local computer. Then I shipped it to DR and attached it to the DR wan accelerator. I mapped the backup and started the job. Every VM from the job where "digested" and synced successfully until it got to the problem VM with the 2TB disk. It failed before transferring anything on the 2TB disk, 5 times and continued processing the rest of the VMs. Tried clearing cache to no avail.

I thought that maybe the seed was faulty so I deleted it. But it wasn't and everything I tried by myself or with support hasn't work. Even a separate job only for that VM fails every time when it gets to processing the disk, 30 seconds in, without transferring any data.

[foggy]

The actual issue is in the disk size - currently the backup copy job cannot handle backup digests for disks larger than 1.8TB (in case of default 1MB block size). The fix is already scheduled for the patch 2.

[VladV]

Thanks Alexander. Actually knowing what the issue is exactly and that it is scheduled for patch #2 for certain is much better than knowing that it was reproduce-able and confirmed and that it will hopefully make it into patch #2.

Thank you,
Vlad