12-05-2011 07:21 AM
our backup first goes to disk, and then to tape.
we have configured 3 backup policies with 6 jobs per policy:
File backup: monday - friday + week/month
Otherservers: monday - friday + week/month
Exchange: monday - friday + week/month
all of the jobs uses one tape and are configured as this:
backup goes to slot [x]
append to media, overwrite if no appendable media is available
enable DirectCopy to tape
the 3 jobs start simultanious at 21:00 in the evening.
the media pool properties are configured as this:
OPP: 3 weeks
APP: 12 hours
sometimes the 3 backup jobs to tape ran fine, but sometimes the first backup goes fine, but the second and third backup not.
the second and third backup are waiting for a free tape. if i move the tape (thas has been written by job 1) to the scratch media pool, the second backup ran fine, and the third backup also ran fine after that.
do we have something configured wrong? or why is this happening.
Solved! Go to Solution.
12-07-2011 12:35 AM
This is because your AP is only 12 hours and the AP period starts from the time the first job writes to the tape. Some of your weekly jobs lasted almost a day, so when they are done, the AP has expired.
12-05-2011 08:26 AM
If you want the jobs to append to the same tape then you will have to start them in a serial way so that one job has finished before the second job starts and not in a parallel way - which will means creating separate policies so that you can start them at different times.
If you start them at the same time then the first job will make a decision on what tape it needs and start using it, the second job won't be able to use that tape as it is busy so will decide on a different tape or switch to expecting an overwritable tape and once it switches to overwrite there is no way back to the append. It will then queue waiting for the drive and tape it has selecetd to be available. The 3rd job will do the same.
However if you let job one finish, it's tape will be appendable and not busy and the second job will be able to use it (assuming your media set properties are OK anyway)
Backup Exec does not support multiplexing (i.e. more than one write to the same tape drive in parallel) - you need NetBackup for that.
I am surprised you ever see your current configuration work - although this might depend on what other tapes are available - I doubt you are getting all 3 jobs on the same tape when it does work.
12-06-2011 11:57 PM
thanks for your answer.
we have configured 3 policies, and the 3 jobs start all at 9:00pm.
daily backups(incremental):
- Exchange, start 9PM, duration 1hour
- Fileserver, start 9PM, duration 4 hours
- Other, start 9PM, duration 5hours
weekly backups(full):
- Exchange, start 9PM, duration 1hour
- Fileserver, start 9PM, duration 23 hours
- Other, start 9PM, duration 22hours
we are using a tape library as destination.
we have configured Partitions for the slots.
the 3 monday jobs will use the "slot 1"
the 3 tuesday jobs will use the "slot 2"
etc..
the 3 week jobs will use the "slot 6"
the daily incremental jobs are going fine to one tape.
only the week jobs has the "waiting for a tape" problem
12-07-2011 12:34 AM
OK is the amount of data in your Full Weekly job actually filling the tape? As your incrementals on the daily's probably aren't which might be why it works for those.
As soon as you target a parttion (so a slot) if the media in that [partition fills up then the job has nowhere to go and will sit waiting for you to insert overwritable media in that specific slot.
I suspect your weekly job may need to go to a partition with at least 2 slots in it (and maybe more)
12-07-2011 12:35 AM
This is because your AP is only 12 hours and the AP period starts from the time the first job writes to the tape. Some of your weekly jobs lasted almost a day, so when they are done, the AP has expired.
12-07-2011 02:51 AM
@Colin.. the backups must fit on 1 tape. this was going fine.
@pkh.. that could be the solution! i have changed the APP to 36hours. i will let you know the status after the weekend.
thanks guys so far
12-15-2011 12:06 AM
changing the APP was the solutions. thanks!