If vm size is more than 500 GB, Normal backup fails!!

[rahulvenu]

The job failed with the following error: The operation failed because the virtual center or ESX server reported a problem creating or acessing one or more files needed. Event ID: 34113.

Can somebody help me out from this issue? I am new to this discussion Forum, and I hope I will find some help.

For your information.

This happens with backup of VMs that are more than 500 GB of size.

In my case I have a windows 2008 r2 OS and Fedora 64 bit OS.

1. Windows system =(80 GB+700 GB) HDD, 4 GB RAM

2. Linux system = 80 GB + 500 GB, 4 GB RAM

It does not matter whether the VMs are online or offline. It does not matter whether the agent is installed or not installed. Error is the same.

My product is Symantec BackupExec 2010.

I went through some kb which explains this issue during incremental/differential backup. My backup is Normal.

[CraigV]

Hi,

I don't think this is a BE issue. We had the same issue on 1 of our sites...Veeam (we use it because we don't license AVVI, and the backups are very fast!) wouldn't snapshot a VM as the *.vmdk file was 512GB + a bit.

End result? Block size on our storage...because it was formatted with a 2MB block size, it wouldn't snapshot correctly. Going forward, we're creating all LUNs with an 8MB block size (which would theoretically allow a 2TB LUN to be used and snapshotted).

You might want to check the block size of that VM...

Thanks!

[rahulvenu]

I have two Harddisk added to the VM. One is 500GB from a DataStore with 8MB block size which is used for storing data and other is 80 GB from a 1MB block size datastore  for OS.

Thank you so much for the response.

[CraigV]

Mmm...8MB block size means you can hqave a maximum size of 2TB.

Is BE fully updated?

[pkh]

Is the backup failing on both the Windows as well as the Linux VM's?  If it is just the Linux VM, then it could be because Fedora is not supported by BE and thus BE is unable to communicate with it properly to take the snapshot.

[rahulvenu]

Both 700 and 500 GB hard disk are in the same datastore with 230 GB of free available space. A BE agent is installed in the Fedora VM, and I am able to take backup of the internal files.

Thank you, CraigV and phk for your response.

[rahulvenu]

Hi,

I tried the VMware Data Recovery(VDR) to take the backup of vmdk files which gave me a similar error. It failed to take a snapshot since the vmx file is in a datastore with 1MB block size.

My OS harddisk is in the datastore with 1MB block size, so the error is ok. Now that I have to find out how the snapshot is related to vmx file as far as a backup software is concerned.

[CraigV · Accepted solution]

...like I said, if it is 512.01GB, it's too big on a 1MB block size, and you'd have to storage vmotion it to a pool formatted with 2MB and bigger. Unfortunately, you might be looking at a complete format on the data store, and if you don't have free disk space elsewhere this becomes a problem.

BE isn't the only application with an issue like this...we run Veeam to back up the VMs entirely, and anything bigger than 512GB fails on the snapshot too. We're in the process of changing this on our MSAs to 8MB block sizes.