Agent Crashing Server?

[kayvonp@re-wa.org]

I had a weird situation with a server recently and wanted to see if someone else had ever seen the same behavior.

Disclaimer -- since it usually gets asked, I do not have a support ticket (yet). I probably won't be able to work on this server for another week or so and didn't want to open the ticket until I was onsite.

Anyway, we have a set of three Windows Server 2008 R2 servers that I wanted to put the Windows Agent on. I had done this on other servers with much success. These servers first needed to be updated to service pack 1 (they are on an isolated network with no outside connection, so they don't get updated often). I did so and then installed the agent. Everything went according to plan. I had to reboot a couple of times, but the agent installed and I was able to successfully run a backup job for all three servers.

Here's where it gets weird.

The next day, I came in to find that none of the scheduled backups had run. I check my console and all three servers are saying they need a reboot. That is weird because I rebooted them two or three times the day before and they were running without that notice when I had left the previous day. Anyway, I rebooted the first machine to see if it would clear the message.

Instead, Windows failed to load. I kept getting to the Windows repair screen. I ended up having to rebuild the boot configuration data (bootrec /rebuildbcd). After that, the OS still failed to load, but it gave me a different error. I really wish I had written it down word for word, but I was trying to hurry to get the machine back online. It failed loading a particular Veeam dll. I think it said something about verifying the file's signature, but I may have the wording off. The machine would not boot into safe mode either, giving the same error. I then tried Last known good config and it finally booted.

I didn't have time to work on it that day, so I went back to our old backup software (Symantec). I had to uninstall Veeam to get it to work.

Anyway, I will probably end up opening a ticket for this, but won't until I actually have access to the machine again. In the mean time, I thought I would check to see if anyone else had ever run into something like this. Any thoughts?

[Rick.Vanover]

I have not - but honestly - nothing surprises me anymore when 2 (or more!) backup products are installed on the same OS. Nothing will go well in that situation.

Why?

There are frameworks that all backup products will use - they use them all differently however. Best example is VSS. It doesn't handle this type of situation well.

Conversely if we are talking VMs - it's very easy to do left-right comparisons of Veeam and another product - simply by not doing the Application-Aware Image Processing. That will not interact with the guest - and specifically for VMware VMs - our backups are all via API - so they are effectively 'no touch' to the source environment.

All of that being said - yes, a case may be in your future as you initially predicted