Windows 2012 R2 Socket Write Error

[AndrewPilgrim]

Hi,

I have an issue with a server that fails to backup the C: drive with an error Status 24: Socket Write Error. Yet the child jobs for the D:, E: & F: drives all work without any problems.

When i start a backup job it looks like its going to start okay. Looking in the detailed status of the job it gets down to where it says "begin writing" and starts to count up the files being written then we get this error;

Critical bpbrm(pid=) from client CLIENT1: FTL - socket write failed

NOTE: I cannot post any log files or outputs on here as these servers operate in our secure cloud environment. So posting logs/hostnames/IP is not allowed. Sorry.

This is a Windows 2012 R2 Virtual Machine and it was built at the same time as another server of the same design which works. We are running 7.5.0.6 client and 7.6.0.4 on the media and master servers.

I have read all sorts of posts about possible network issues or client timeouts but since this server with the problem is sat on the same ESX host as the second server built at the same time (which works perfectly) i cannot see networks this being the issue here.

So i was wondering if anyone might have any ideas where to start looking? 

[Marianne]

So - backup of a VM with NBU client installed? 

Any reason why you are using normal client-based backup and not VMware policy?

Have you tried to enable bpbkar log with level 3 logging?
You should be able to see in the log if the backup is getting stuck on a particular file or folder.

How long does it take for the job to fail?
Have you checked Client Read Timeout on the media server and KeepAlive settings on client and media server?

Have you checked bpbrm and bptm logs on the media server? (Also level 3). 
bptm will show up to when data is received from the client.
bpbrm will show metadata received from client. 

[AndrewPilgrim · Accepted solution]

Hi Marianne,

Thanks for the reply.

We've just this moment found the solution. It would appear its related to the ESX host after all. We ran another set of tests and the server that was working, started to fail which made us wonder if the ESX hosts is overloaded. And the server that did work only worked more out of luck. We migrated the VM to another ESX host and the test backups started to work okay.

So going to close this off. Thanks for replying anyway.