cancel
Showing results for 
Search instead for 
Did you mean: 

BE 2010R2; how to get job to continue if 1 server in selection list fails..

Edward_Charles
Level 6

We have a new Win 2008 server with BE 2010R2 backing up 5 other servers, a mix of 2003 and 2008 incl. Exchange 2003 and 2010 all in 1 selection list.

The question is the 2010 Exchange server failed to backup properly because of some permission issue (Backup Exec attempted to back up an Exchange database according to the job settings. The database was not found, however. Update the selection list and run the job) however the real question is why would it not just continue the job just skipping over the server that it cannot backup in the selection list and still backup the remaining servers?

This server was even the second last one in the selection resource order list but when starting the job it instantly fails in 2 seconds.

Our experience is that the jobs usually continue backing up the servers that it can and at the end it just becomes "failed" if it missed a server or file or something.

Any thoughts on this would be greatly appreciated.

Thank you.

Regards,

 

8 REPLIES 8

Edward_Charles
Level 6

Does anyone have any thoughts on this please?

Is the job supposed to skip a server if it fails?

Thanks

Colin_Weaver
Moderator
Moderator
Employee Accredited Certified

I think it depends on the type of failure.

i.e. I have a feeling that a failure where the remote server does not even respond (because it is offline or non-contactable) might skip to next server in list, but an actual failure where the remote server does at least initially respond might cause the job to stop.

Note this is a guess - and if I am correct, I doubt anyone could list what types of failure will cause the job to stop and what types would cause a skip of the affected server.

Smitty42
Level 3

I have had similar problems before when backing up multiple servers in a single job.  I simply broke each server up into it's own job/policy and scheduled the jobs to start in the order i want.  i still have one 2003 server with SQL on it that fails occasionally and i have it scheduled last so it doesn't keep the media occupied and stall other jobs.

teiva-boy
Level 6

Best practices dictate that you should break up jobs based on applications, OS's, and further break up jobs down to inidividual servers (if allowable).

At the minimum, Exchange SHOULD be on it's own job/policy.  

Then the 2003 servers in their own, and 2008 in their own.

Ideally, you would have a single job for each server, though Exchange still needs to be on its own.

 

Why?

Maintenance is easier.  Need to take down a server, well it wont affect an entire job.  Have a downed server or a failed server?  Well again, it wont take down an entire job.  Certain applications have very specific settings that only apply to it, and not an OS.  Thus it's recommended you break it out to its own job.  

Edward_Charles
Level 6

Thank you all for the replies.

THe reason I prefer to have multiple servers in 1 selection lists as I find it easier to manage the jobs. Instead of managing 10 jobs for 10 servers with different starting times and durations (without overlapping jobs) you could use the 1 job backing up those 10 servers providing it'll just skip over a server if one has an issue which it usually does. I see the point however for server maintenance as you cannot restart the servers while the backup job is running.

Thoughts?

THank you.!

Regards,

teiva-boy
Level 6

100% backup failure, or a 90% success rate...  In your example, that is what you would have if you continue the way you backup.  Just saying...

 

Ease of management is subjective, and you are looking at it from a perspective running a single job.  When in larger environments, we look into it from a perspective of successful backups, and what makes management happy (or auditors too)

 

Edward_Charles
Level 6

Thanks for your comments!

Our experience is however that even if 1 server is down or something failed on it the rest of the backup and servers complete successfully even if the completed status might say "failed". Our environments are anywhere between 2 to 10 servers at a site. We usually backup the Exchange servers and file servers separate and it also depends on how much data they have. One client has over 1TB of data and then we split it up as sometimes there just aren't enough hours in a night...

In your case do you create a different destination b2d device (a LAN folder) for each job and each server then?

Thank you for your input.

Regards,

teiva-boy
Level 6

If you are using a B2D, with enough performance, you can send multiple jobs to that single target concurrently.  You dodn't need to create individual B2D targets if you dodnt have a logical or performance reason to do so.  That is the benefit to B2D's.  Where as tape can only dod a single job at a time.  

By using more concurrent streams, you can shrink backup windows.  :)