I've got some questions about Exchange 2003 SP2 backup because our backups are taking way too long (2+ days). Today, I use ntbackup and take three full backups per week. There are three backup jobs scheduled and each job backs-up all of the storage groups on the server. The bkf files are ~400GB. To shorten the backup duration, I'd like to create a separate backup job for each storage group and schedule them to run simulatneously (maybe starting within 10 minutes of each other). My questions are, will that work and what implications will it have for the total size of the backups? I'd also like to run a differential backup in between these full backups. What will happen if a differential backup starts before the full backup finishes? Thanks

2 days? I would try to figure out why its taking that long before doing anything else. Is the backup server up to the task? Where is the bottleneck?

I'm working on determining that but am still interested in the possibility of running three separate backup jobs concurrently and/or running a full and diff overlapping. Our next backup starts this evening so I'm going to setup perfmon to watch disk and network on both the Exchange and backup servers. I guess a third question in this thread could be, "any recommendations for the proper counters?" I haven't even looked yet.

How large are the databases? What kind of disk are you backing up to? -- Ed Crowley MVP "There are seldom good technological solutions to behavioral problems." . "mhashemi" wrote in message news:ae625588-b398-4234-81e7-5215485c4e5d... I'm working on determining that but am still interested in the possibility of running three separate backup jobs concurrently and/or running a full and diff overlapping. Our next backup starts this evening so I'm going to setup perfmon to watch disk and network on both the Exchange and backup servers. I guess a third question in this thread could be, "any recommendations for the proper counters?" I haven't even looked yet. Ed Crowley MVP "There are seldom good technological solutions to behavioral problems."

The backup array is full of SAS drives and I'm connecting to a Windows share over a gig-e network. The databases are 88GB, 171GB, 95GB, and 56GB.

hmmm, this reminds of an old problem with ntbackup. Microsoft was running into essentially the same issue. Ignore the cluster server part and look at the settings you can adjust to increase throughput - (ntbackup with the /fu switch) http://technet.microsoft.com/en-us/library/bb735157.aspx Backup Process Used with Clustered Exchange Server 2003 Servers at Microsoft http://support.microsoft.com/kb/839272

I tried adding the /fu switch to our job, but it didn't help. I checked the version numbers for the ntbackup hotfix and my version of ntbackup and mine is newer (we're running SP2).

Crickets?

Did you make the registry changes as well?

Right, about that...since we're not using a tape drive, what value should I enter for the HKEY_CURRENT_USER\Software\Microsoft\Ntbackup\Backup Engine\Max Num Tape Buffers? Does it matter?

This must be a toughy :)