I'm posting this in case anyone else has this problem. It's been reported to Veritas, but isn't on the knowledgebase yet. A few weeks ago, I upgraded to version 10 and found that after my first weekends full backup, BackupExec had changed the schedule on several jobs, causing them to run twice.
There had been no intervention by myself. In fact the backup server was completely unattended all weekend. I am still waiting for Veritas still have to reproduce this, or confirm the cause, but I believe other people may well be affected by this problem so am reporting it here.
I am confident that I know the cause of this problem:
In BackupExec 9, I created several jobs from a single template. Later on, I edited one of these jobs individually to change it's schedule. This broke the link between this job and the template since they were now running on different schedules.
When I upgraded to version 10, BackupExec created policies for all my old templates. What I did not notice at the time was that this policy was linked to ALL of the jobs originally created by the template, including the job that should have been kept separate as it had a different schedule. I did not notice anything unusual however since the upgrade had retained the scheduled times for all the jobs. It was not until that jobs ran and BackupExec processed their next scheduled date that the problems started.
It appears that when the job with a different schedule ran, BackupExec updated the schedule of the entire policy to match the schedule of this job. This change happened automatically and immediately, and caused all the other jobs to run immediately this job finished.
I found that Version 10 has no way for me to individually edit a job, so I could not change the schedule of this one job while it was within the policy. My only option was to now create a new policy for this job. However, I found that even deleting this job and creating a new policy for it did not fix this problem. I actually ended up with two policies exhibiting a bug which caused the schedules on both to match, regardless of which one I edited.
The only suggestion from Veritas support has been to delete and re-create all my existing policies and jobs. I believe this is going to fix the problem, but it's going to be a week or so before I can know for sure.
As I said, this hasn't been reproduced by Veritas yet. However, the problem definately exists, and I am pretty certain (at least 95% certain) that the steps above are exactly what caused it.
If you believe you may be affected by this, the problem can be spotted by comparing the scheduled run times for your jobs shown in the Jobs List to the run times shown in the Calendar view. The Jobs List displays the next scheduled run time for each job whereas the Calendar uses the date from the policy. Any difference between these two views indicates that you have a policy running on a different schedule to a job and you are likely to have serious problems unless you take action to correct this.
I hope this helps a few people. Please post a reply to this message if you were affected by this bug (I'm curious about how many people it could hit).
Ross
