Forum Discussion

PXCSW's avatar
PXCSW
Level 3
7 years ago

V-79-57344-38264 - Unable to copy the virtual machine disk using the VMware VixDiskLib

Hi,

we've been having issues with failing Backups of VMs in our VMware infrastructure lately.

The error get is this:

V-79-57344-38264 - Unable to copy the virtual machine disk using the VMware VixDiskLib

VixDiskLib_Read() reported the error: Memory allocation failed. Out of memory.


We run Backup Exec 16. I recently installed FP2. We had the issue before FP2, but only maybe 1 Backup per day, that failed. Now its rather 3-4 per day. Both incremental and Full Backups.

The "out of memory" Error doesn't make much sense to me.

The Media Server runs Windows Server 2012R2 with 32GB of RAM. According to monitoring it never dips below 70% free RAM. Our ESX Hosts also have multiple hundred GB RAM free at any time. The VMs themselves also don't show any RAM usage issues. The vSphere Appliance has 16GB RAM, usually at less than 2GB used.
VMware vSphere & ESX Version: 6.5 with latest updates

Backed-Up VMs run Windows Server 2012R2 / 2016, some Linux

  • Is there RDM configured in those VMs? Pleae have a look at this article.

    If that is not the case, please try to isolate the issue, by doing below. You may will get some clue after isolation.

    1) Create a separate job for the VMs that are getting failed and run it to a different type of storage.

    2) Run non-GRT backup and see whether that works.

    • PXCSW's avatar
      PXCSW
      Level 3

      There is no RDM configured for any VMs.

      Every VM is stored as vmdk on a VMFS 6 Datastore and Disks are in "Dependent" mode.

      To make things clear:

      The issue isn't limited to a couple of VMs. It happens on all of them, occasionally. Some Jobs will work for weeks without any issue, then fail one day, sometimes 2 in a row, then work the next day. I had 3 out of 26 virtual backups fail between yesterday and today.

      Every VM has its own Backup Job. All backups go to one of two disk based storages, it doesn't seem to matter which one.

      Also forget to mention, they all seem to fail at the END of the Backup. On a Full Backup the failed ones show pretty much the same runtime / bytes copied as the succeeded ones.

      • PXCSW's avatar
        PXCSW
        Level 3

        It looks like I solved it. Haven't had any failed backups over the last 2 days.

        The issue seemed to have been a registry change I made back in February of 2016.

        I set values for SAN/10GB NBD according to this article:

        https://www.veritas.com/support/en_US/article.TECH185691

        But couldn't see any difference in speed, so I bumped the Size of Read Buffers and Size of Write Buffers keys up to 16392 a month later. Again couldn't see much change, but it didn't have any negative impact at the time, so I left it.

        Dropping it back down to 8196 and restarting BE services seems to alleviate my current issue though. 

        Still weird how this was OK for almost 2 years.