Horribly slow restores from incremental backups

[awsag]

Hello

We have a backup location where we restore our Live Servers for hot standby.

We used Vranger before and now with backup exec the restore process for some VM takes up a huge amount of time... I have uploaded the logs for the worst case.

The full backup took 24 minutes 12 seconds and 52.9 GB of data backup up.

The first incremental backup took 7 minutes 22 seconds with 13.2 GB data.

The second incremental backup took 25 minutes 51 seconds (!!!) with 80 GB data (!!!!). How is this possible? An incremental backup which uses up more space than a full backup and takes longer?

The third incremental backup took 3 minutes 26 seconds with 4 GB data.

The restore from the third incremental took 2 hours and 58 minutes and it shows 150 GB of data restored which is even larger than the hard disk of the VM. How is this possible?

Backups were taken every week.

All the servers which take that long to restore have such a strange incremental backup in them where the data backuped shown is equal to the size of the disk.

I find it strange that incremental backups do take up sooooo much space for most servers. Our citrix servers where the full backup takes up between 70 and 90 GB shown the incrementals show between 10 and 90 GB...

Edit: I just did a restore from the fullbackup (server from example above) and it took 1 hour and 30 minutes. Why does this take 3 times longer than the backup? It's not the infrastructure for sure because other older Windows 2003 VM which take 5-6 minutes to back up also only take 6 minutes to restore with 3 times the throughput... The server from the example is a 2008 R2.

[awsag]

No one has an answer or explanation to this?

It's really getting frustrating working with this new Backup Exec with the pletora of problems we have :(

I just restored 2 servers i backed up on the last weekend with a full backup. Backup time 20 minutes, restore time 1 hour 24 minutes... I understand restoring taking a bit longer but how can a backup have 4 times the throughput than the restore???

[VJware]

Full backup backed up 56'762'023'582 bytes and wrote to B2D000655
Incremental backup backed up 85'899'379'283 bytes and wrote to B2D001701
However, your restore took time because it was restoring from B2D002346, B2D001701, B2D001176 & B2D000655. And per the job log, it shows more data being restored and hence more time was taken. I am guessing it is because there were additional incrementals (as you can see it restored data from other 2 B2Ds as well) which were part of one recovery chain.

And as to why more data was backed during incrementals, there can be multiple causes. I have listed few of them below:-
1) Maybe CBT isn't functioning properly. Try to re-enable CBT and this KB may help in doing so - http://www.symantec.com/business/support/index?page=content&id=TECH129452
2) If you are using virtual RDM disks for the VM as incremental/differential backups aren't supported for this type of disks.
3) If using ESX 5.1, have a look @ this KB - http://www.symantec.com/business/support/index?page=content&id=TECH223025
If none of these are applicable, debugging the incremental backup would help us find out why more data is being backed up. Would recommend to log a formal support case.

[awsag]

Thanks for the answer, we are indeed using ESX 5.1 and i will have to take a look at that.