SureBackup Job Linux Docker Host Ping Fail

[mihlfeld-it]

We have a SureBackup job setup that seems to be working well overall running through firing up a lab environment and doing basic tests of our backups on said lab environment. We are having a bit of a problem with one of our VMs during testing, however. The VM is an Ubuntu 18.04 server that is a docker host. For some reason beyond my understanding, SureBackup is latching onto and trying to do a ping test against one of the docker virtual networks on the virtual machine and isn't attempting to test the actual VMware assigned NIC (just a single NIC) when doing ping testing. We have two other Ubuntu 18.04 docker hosts on our network that don't have this issue at all. Ping tests are against the real IP of the host and not a virtual address setup and used by docker.

Any clues / ideas as to why it's doing this on this one host, but isn't on a couple of the others? I can do an instant recovery of this machine and any ping test to the actual VM NIC that is accessible on the network will work just fine.

[HannesK]

Hello,
but you can see the "correct" IP in VCenter? I'm just guessing that the two other working hosts might have a different order of IP addresses in VCenter (which shows the IP addresses which they get from VMware tools)

Best regards,
Hannes

[Andreas Neufert]

We need to check the logs.
But overall we receive from vcenter the list of IP addresses and use them to ping the interfaces. Maybe it is pure luck or specific IP order within the vcenter VM list of IPs that we use the correct one.

[PetrM]

Just one more addition:
Ping test is performed against every IP address which exists in an array of IP addresses returned by VMware tools running on a VM in SureBackup once stabilization point is reached (look under "Stabilization Algorithm"). Probably, there was just one IP address returned and this IP belongs to a docker network adapter.

Do you see all IP addresses in vSphere client when you run instant recovery of the VM?

Anyway, I'd suggest to contact our support team in order to analyze logs and to check all versions mentioned above.
For example, they could help you to open a case with VMware support as well if there was just one IP returned.

Thanks!

[mihlfeld-it]

I am able to see the correct IP in vCenter, both for the machine as it's running normally and when doing an instant recovery. However, I just did an instant recovery of the machine and I noticed it took a bit before VMware tools got up and running and started reporting IP addresses. The first in the list is the Docker IP that invariably seems to be the one getting pinged. The actual machine IP is 3rd in the list (2nd is some IPV6 IP).

I did just do an instant restore of one of the other machines that is running docker (but is successfully passing the tests from SureBackup) and it basically was the same behavior. It took a bit before VMware tools started reporting in and then the actual machine IP was listed, but it was about 5th down in the list behind a couple of the docker IPs and IPV6 IPs.

[PetrM]

Ok, I'd say that for this particular VM, the list of IP shown in vSphere client is different than a list of IP returned to our specific API request, most probably we receive the single IP.
I'd suggest to increase "maximum allowed boot time" startup parameter and to contact our support team if it doesn't help, just please don't forget to paste case id for our reference.

Thanks!