I am runnning a NetBackup 7.6.0.4 environment. My media server/backup host is Red Hat Linux 6.5, and vCenter is 5.5. I am using a query-based vmware policy to backup my virtual servers.
I recently came across an issue where all of my VM backups were failing and the error code was in the job details was:
Error nbpem (pid=12345) No server credentials configured for [vSphere]
Operation status: 4238
I checked the vmware credentials in the admin console, and no issues there. I scoured the forums, and made sure DNS resolution (forward and reverse) was as expected, and checked firewall to make sure backup host was allowed to communicate via port 443.
After troubleshooting with a Veritas engineer for over an hour, he had me run vxlogview to generate some log info. I naturally piped it to /tmp, to discover that /tmp was already 100% full. That was causing the issue. After cleaning up /tmp of some backup-related vmware files, the problem went away.
I had a similar issue in a 7.5.0.6 domain, after increasing the vmware backup thresholds, though the errors NBU returned there was error code 130, and/or 71.
The standard build for our RH servers has /tmp in swap, and only 1GB. According to Veritas there is no way to redirect NB's use of /tmp for vmware backups, so the only fix is to create a filesystem for /tmp, and make it > 1GB.
Hope this helps.
