04-01-2012 01:06 AM
Hiya,
I've got about 6 jobs running everynight at various times so they don't overlap each other. These all normally complete by midnight but yesterday morning I kicked off a Backup/Duplicate job of a new Exchange server. This is then duplicated to our other site on completion and due to the large initial backup (we use dedupe), the duplicate job is still running (about 105GB of 120GB total).
Well this duplicate job has caused all other jobs to pretty much grind to a halt. For example our Web server job usually takes 3m15s at a rate of over 500MB per minute to backup 1.5GB. Now it's running at nearly 11h for just over 1GB at a rate of 1.00MB per minute.
What's going on. Does a duplicate job bring all other jobs down to its rate? The duplicate job is over a 10Mb WAN link.
Thanks
Windows Server 2008 R2 SP1, BE2010 SP2 + hotfixes, 4x 2TB disks in RAID 10 for the dedupe folder.
04-01-2012 01:23 AM
When you duplicate from a dedup folder, the data will have to be hydrated first and this takes time and horsepower. Even if you are using optimised dedup, the data will be hydrated first with BE 2010 because your Exchange backup is GRT.
When you use BE 2012 and optimised dedup, then GRT data which is dedup'ed will not be hydrated before they are duplicated.
04-02-2012 12:51 AM
Ah OK, so that would explain the initial time it takes to do the initial duplice job on a clean install of Backup Exec (the hydration thing). We are duplicating to another de-dupe folder, but we're not using GRT for Exchange due to the added time it takes to complete (about 1h vs 12h+!!).
We're actually already running 2010 R3 with SP2 + hotfixes. I missed out 'R3' in my original post.