Good afternoon,
I was wondering if someone could help me with this weird behavior, which is only partially related to NetBackup.
NetBackup for VMware snapshots failed with status 40. They worked just fine the other day. Snapshots that would else be finished within 20 second now took up to half an hour before failing. Every now and then a snapshot would succeed. But the majority of the VMware backup jobs failed.
Of course nobody changed anything.
After an extensive and not particularly successful troubleshooting session my customer finally happened to remember that he had indeed changed a little something. He had added a proxy server to in the SLES system configuration.
Now, the problem lies in the way NetBackup handles it’s communication with the virtual center server _and_ in the way SLES uses it’s proxy server’s whitelist.
Let’s say, we have a whitelist containing two entries:
- vcenter01.customer.de
- vcenter01
Checking with “wget” we see that connections to the FQDN (vcenter01.customer.de) bypass the proxy server, while connections to the short name (vcenter01) will always be routed through the proxy.
Obviously at some point in the snapshot creation process NetBackup uses the short name and gets stuck with the proxy server. A reverse lookup of the vcenter IP only returns the FQDN. So that would be no starting point for a workaround.
Could anyone think of a solution for this dilemma?
Many thanks,
Thomas