RHEL 7 & SureBackup

[stewsie]

Hi

Is anyone using SureBackup to test backups containing RHEL 7 VMs? I have a problem where during the SureBackup job a new adapter is added and it is ignored by the OS. The adapter is an E1000. I had a support call open with Veeam that resolved this for RHEL 6 but I had to close that call due to problems at our end.

I would really like to run SureBackup jobs against backups containing RHEL 7 VMs.

[foggy]

Paul, would you mind to share the support case number?

[stewsie]

This is the one I have open at the moment 02150128

Still struggling to make this work which means I am unable to test any backups with RHEL 7 VMs

[nitramd]

Paul,

VMXNET3 is the preferred driver; see https://www.lewan.com/blog/choosing-the ... he-vmxnet3.

Hope this helps.

-MD

[stewsie]

We found that for the Linux VMs that the E1000 gives us the least problems especially regarding the VM Tools. This is unrelated to the SureBackup issues I have though.

In fact with the VMXNET3 adapter it introduces a similar problem in Windows where a new adapter is used during SureBackup and an IP is not allocated. There is supposedly a hot fix from Microsoft to resolve that. https://www.veeam.com/kb1570

Cheers

[lightsout]

Sorry to drag this back up, but was it ever fixed?

I'm trying to do a Sure Backup with RHEL 7 and Veeam 9.5 U3a, and it doesn't ping, but RedHat ignores the NIC. We're using DHCP still. I know this is not technically a Veeam issue but a RedHat one, but I imagine it is a common issue!

[stewsie]

Still waiting to do more testing with the Linux experts

[nitramd]

I can confirm that CentOS 7.3 has the same issue - no surprise.

I'm working on creating a test Ubuntu machine to see what happens.

[Mike Resseler]

Hi all,

Indeed it seems to be something that can be reproduced without Veeam so "technically", as Paul said, it is not a Veeam issue. That being said, we are eager to know the solution also so Paul, please keep us informed. @nitramd, please keep us also updated with your tests.

[nitramd]

Ubuntu 18.04 LTS with latest updates:
IP address hard coded
Backup successful
NIC present in SureBackup environment
Heartbeat & ping tests successful

CentOS 7.5 with latest updates:
IP address hard coded
Backup successful
NIC present in SureBackup environment, although running command ip a s showed NIC up but no IP address
SureBackup job failed with message: Error: Results: OS did not boot in the allotted time

It seems that some progress has been made in version 7.5 with the NIC being present but I have no idea why the IP address did not show up in IP a s. I don't know the status of the other versions below 7.5, except 7.3.

I would suggest, if possible, logging into access.redhat.com to check any KB articles or opening a ticket. Otherwise, try the CentOS forums to see if there's any information there.

[nitramd]

Neglected to mention NIC detail

Ubuntu 18.04 LTS configured with an VMXNET 3 adapter

CentOS 7.5 configured with an E1000 adapter

[Mike Resseler]

Interesting.

Any change you can do a test on the CentOS 7.5 with an VMXNET3 adapter?

[nitramd]

Mike,

That's next on the test list - I ran out of time yesterday.

[nitramd]

CentOS 7.5 with VMXNET 3 adapter:
IP address hard coded; adapter name ens160; configuration file: ifcfg-ens160
Backup successful
NIC present in SureBackup environment, NIC ens160 present and has a status of up but no IP address

Poking around in the SureBackup environment, I eventually noticed that the MAC address of the NIC differs from the VM, i.e. original NIC MAC address xx:xx:xx:xx:79:6c, and in the Sure backup environment the MAC address is xx:xx:xx:xx:23:9A with the ifcfg-ens160 file missing (deleted?)

Strangeness abounds.

[Mike Resseler]

If I understand this correctly, when you logon to the machine running in the surebackup job, you cannot retrieve that file that was originally there during the backup?

[nitramd]

The file ifcfg-ens160 is there, I missed it the first time around. Mea culpa.

I've brought up the CentOS test twice in the SB environment twice today. When I check the ens160 adapter, there's no IP address and the MAC address differs each time the machine is brought up in the SB environment.

Certainly not a Veeam issue.

[lightsout]

Mostly correct - the problem is Veeam doesn't see the VM booted up, and shuts it down. So I'm unable to use a Surebackup job for a test lab, as I can't use any RHEL 7 machines.

[lightsout]

So my issue is described here:

https://access.redhat.com/solutions/2592561

My fix - I created a bogus ifcfg-ens192 file in /etc/sysconfig/network-scripts without MAC info, and then it worked.

[nitramd]

@lightsout
Interesting fix, thanks for posting.

For my CentOS 7.5 machine, I commented out the MAC address reference in the ifcfg-ens160 file, ran a backup, then brought up the machine in the SB environment. The hard coded IP address was present (same the as the VM's IP) and the MAC address was different. I noticed that vSphere had an alarm that stated the machine in the SB environment had the same MAC address as the machine in the VM environment.

The ping test failed.

[lightsout]

So for me, I have ifcfg-eno18021 or something like that.

I now have a 2nd ifcfg file called ifcfg-ens192, as when the SureBackup starts up, it always comes up with that new name (per the RH link).