Kaseya Community

Alert on Non-Action

  • Ok, not all our clients use the Kaseya Backup, so we montior the event logs for Retrospect and Backup Exec to confirm success or failures.

    The problem is, sometimes these programs don't run, the services are running, but the scheduled event isn't. (In Retrospect there is a STOP on the toolbar that prevents any scheduled events from running.) In these cases, we simply don't get any alert at all.

    Is there a way to alert us when a known event "doesn't" happen. In other words, if the backup didn't run last night, can we get an alert to that fact.

    Since we get so many alerts and success and failures, its hard to know that we didn't get an alert from a client we are suppose to get one from?




    Legacy Forum Name: Alert on Non-Action,
    Legacy Posted By Username: nachnet
  • outofdate.txt
    >> Is there a way to alert us when a known event "doesn't" happen.
    >> In other words, if the backup didn't run last night, can we get an alert to that fact.

    Not built-in to Kaseya, but there is a possibleworkaround.

    Many infrastructure applications write a log of some sort (ie backup software), or need a regularlyupdated pattern file (ie anti-virus). The timestamp on this file can be used to detect the absence of an event, ie backup didn't run, anti-virus patterns were not updated, etc.

    Use the attached script to alert when the file is out-of-date.


    Legacy Forum Name: Backup and Disaster Recovery,
    Legacy Posted By Username: jlawrence@databalance.com
  • I've had a project planned for quite some time now that I just haven't had time to finish that deals with this exact issue.

    Like you said, sometimes, a backup just won't run without giving an error message. Sitting and asking for another tape or drive is sometimes a cause; sometimes a process that just craps out but doesn't actually stop is another reason. Or sometimes a paused execution or corrupted schedule can be another reason (if a schedule is corrupted, for instance, there won't necessarily be any error).

    The whole concept of the program I'm working on is that you'd configure the program with a separate data record for each backup that's supposed to run. You'd configure various parsing parameters - like the client, the backup set name, the product, etc. You'd configure how often it is supposed to run and the allowable number of errors.

    On the backup side of things, you'd set them all to alert when the jobs are completed and to send that to a specific mailbox.

    The program (that I'm writing) then would sit and run on a machine somewhere 24x7. Periodically, it would pull all e-mail from the configured mailbox via POP3. It would then parse those messages, compare against the information that you have on file, and act accordingly.

    For instance, if it can detect that a backup completed for a job named "Default" for server JSMITH1, it will then "mark" in the database that it had completed and all is well. It will also check the errors in the e-mail alert to make sure that there aren't more than you'd normally expect (because most backup jobs seem to have a certain number of errors, like open files that it can't backup or files that change rapidly and don't compare properly).

    If the backups run within the preset time interval (maybe you'd set it for 25 hours), then nothing happens. But if a backup does NOT run properly within that time frame, then it will send out an e-mail alert letting you know that something didn't work.

    Basically, it makes it a lot easier to keep track of what ran and what didn't without having to have a list that you manually check every day.

    When will I be done with this? Hard to say but like I said, I've been sitting on it for probably a year now because of other projects I'm working on. Maybe in a month... maybe.




    Legacy Forum Name: Backup and Disaster Recovery,
    Legacy Posted By Username: warever
  • Most backup programs have the capability to perform a Pre and Post command before or after the scheduled backup runs. You could have a Pre command create a text file before the backup runs.

    You could then run a script a few minutes after the backup should have ran and look for this text file. Depending on if the file existed or not, you could send an email, create an application event log using the Eventcreate utility, delete the text file, etc.

    If using the Eventcreate utility, you could setup a monitor that looks for the application log event you created and alert/alarm as required.

    You could also use a Post command (or both) to create another text file and then perform the same type of scripting check as stated above after the backup was scheduled to run.

    (Eventcreate is a native command in 2003, XP and Vista).



    Legacy Forum Name: Backup and Disaster Recovery,
    Legacy Posted By Username: gssloan
  • gssloan..

    Your suggestion depends on the backup service actually running.. what do you do if the service has stopped (with no log entries)...

    Or an event was missed for say insert tape and now the backup is waiting and waiting for a tape to be inserted... the software could wait for weekswaiting for a tape, which has the effect of suspending backups until the tape is inserted or the job is canceled. In this case there will be no more events logged until the stalled job is cleared.

    hc




    Legacy Forum Name: Backup and Disaster Recovery,
    Legacy Posted By Username: howardc
  • howardc,

    1). As far as the backup service not running, it doesn't matter as far as the solution is concerned. That's why you check for the existance of the text file. If it's not there, the backup didn't run (assuming in your previous check, you deleted the text file after you had verified its existance).

    2). Waiting on tape - this is when you check for the existance of the text file that is created Post backup. If it's not there, either the backup never completed, or it is still running. (Once again, assuming in your previous check, you deleted the Post text file after you verified its existance).

    howardc wrote:



    gssloan..

    Your suggestion depends on the backup service actually running.. what do you do if the service has stopped (with no log entries)...

    Or an event was missed for say insert tape and now the backup is waiting and waiting for a tape to be inserted... the software could wait for weekswaiting for a tape, which has the effect of suspending backups until the tape is inserted or the job is canceled. In this case there will be no more events logged until the stalled job is cleared.

    hc





    Legacy Forum Name: Backup and Disaster Recovery,
    Legacy Posted By Username: gssloan