Forum Discussion

Artegic's avatar
Artegic
Level 6
11 years ago

Monitoring Backup Exec

Due to the idiosyncrasies of Backup Exec 2012, my daily backup run consists of twelve jobs: seven scheduled ones and five linked duplication jobs. Any of these can produce one of several results:

  • completed successfully
  • completed with inconsequential warnings (eg. "new VM added")
  • completed with warnings that require action (eg. GRT failure)
  • completed with inconsequential errors (eg. tried to back up non-existent file)
  • completed with errors that require action (eg. the various internal errors BE occasionally emits)
  • waiting for operator intervention
  • hanging for no apparent reason
  • never started

I need a way to inform my operators whether last night's run completed successfully or any intervention is necessary. Right now they have to connect to the media server every morning, fire up the Backup Exec console, go through the job log checking off all the jobs that did complete, and then click through all the jobs that didn't complete to check whether they started and in which state they are.

Is there a better way?

 

  • Kind of difficult for us to know if an error for tried to backup a non-existent file is important or inconsequential as you might be backing up certain files for a reason and if one of those files then goes missing you would need an alert.

    If you are referring to the System State condition, that some customers see, where registered drivers etc have gone missing, then that is the setting within the Operating System registry telling Backup Exec that something is important that is no longer present. Again we have no way to tell if the missing file is important to decide whether or not an alert is needed.

    As such it would be very difficult to not need human interaction to make that decision and equally difficult for monitoring software to make the same descision unless the software had some kind of user definable ignore list that you could build up over time.

     

    I however would recommend both the Storage Online services mentioned by Guillaume and also updating to Backup Exec 2014 when it is released to get back the job monitor,

7 Replies

  • You can use BEMCLI to give you a nice report p of the job history. See my article below https://www-secure.symantec.com/connect/articles/using-bemcli-speed-tasks
  • Unfortunately there is no way of generating an E-mail notification for a job that just hangs or wasn't started.

    I'll look into the BEMCLI approach.

    Thanks,
    Tilman

  • Hello,

    you could use a tool of storage online called online monitor for BE working on the notification per email and having central reporting.

    http://www.onlinebackupmonitor.com/

    hope this can help.

  • Kind of difficult for us to know if an error for tried to backup a non-existent file is important or inconsequential as you might be backing up certain files for a reason and if one of those files then goes missing you would need an alert.

    If you are referring to the System State condition, that some customers see, where registered drivers etc have gone missing, then that is the setting within the Operating System registry telling Backup Exec that something is important that is no longer present. Again we have no way to tell if the missing file is important to decide whether or not an alert is needed.

    As such it would be very difficult to not need human interaction to make that decision and equally difficult for monitoring software to make the same descision unless the software had some kind of user definable ignore list that you could build up over time.

     

    I however would recommend both the Storage Online services mentioned by Guillaume and also updating to Backup Exec 2014 when it is released to get back the job monitor,

  • Looks quite interesting, though it might be a bit oversized for our needs. I'll definitely have a look.

  • Colin, you're absolutely right of course. The final decision whether or not intervention is necessary will always need to be made by a human. What I'm looking for is a tool to help him or her to do that efficiently.

    One aspect of that is to avoid polling. Currently my operator must actively check Backup Exec at least once a day to see whether everything is well. That takes time, is easily forgotten in the heat of the fight, and regarded as a tedious chore. A push notification of some sort when something deviates from the norm would be nice. False positives (notifications sent when there's really no need for intervention) aren't a big problem as long as they are not too frequent. False negatives (no notification sent even though there is a need for intervention) do however render the mechanism useless. That's why Backup Exec's E-mail notifications do not help me.

    Another aspect is making the decision itself more efficient. Actively checking would be ok if it consisted of a glance at a dashboard to see everything is green or the bar chart looks the same as always. But it takes more than a minute to start the console application and get to the job history, and twenty seconds to open a job log in order to investigate a "failed" or "completed with exceptions" status. (Actual times from a dual six core Xeon with 16 GB RAM in case you wondered.)

    My point about the non-existent files refers to the regular errors I see about temporary files that Backup Exec apparently sees during its scan phase and loudly complains that they are gone once it gets around to actually backing them up. Just like the point about GRT warnings it served only as an illustration that Backup Exec's distinction between "failed" and "completed with exceptions" on its own isn't enough to decide whether or not an intervention is necessary.

    The Storage Online tool looks promising and from all I read here Backup Exec 2014 will be something to look forward to. I do hope the migration won't present too many pitfalls.