Unsolved
This post is more than 5 years old
2 Intern
•
2K Posts
0
552
January 5th, 2007 08:00
Slow Exchange backup on AutoStart Cluster
Hi,
I am using AutoStart 5.1.2 with Exchange module.
I have configured NW 7.2 Build 172 and Exchange module 4.1.
When we backup manually from either physical node, it backs up very slow - 25 GB took 11 hours. When we run server initiated backup for the virtual node, it smoothly backs up one Storage group (5GB) but for the second(25GB) - it starts off well but then backup speed is OK easily around 10-12 MB/sec and backup finishes in around 2 hours.
The problem is that the users get disconnected at this time and get reconnected only when either the backup is over or is terminated.
In a way, Exchange server is not usable during backups. Both physical nodes are 4 GB Dell machines with Windows 2003 & Exchange 2003 so there should not be a resource crunch.
Any suggestions.
I am using AutoStart 5.1.2 with Exchange module.
I have configured NW 7.2 Build 172 and Exchange module 4.1.
When we backup manually from either physical node, it backs up very slow - 25 GB took 11 hours. When we run server initiated backup for the virtual node, it smoothly backs up one Storage group (5GB) but for the second(25GB) - it starts off well but then backup speed is OK easily around 10-12 MB/sec and backup finishes in around 2 hours.
The problem is that the users get disconnected at this time and get reconnected only when either the backup is over or is terminated.
In a way, Exchange server is not usable during backups. Both physical nodes are 4 GB Dell machines with Windows 2003 & Exchange 2003 so there should not be a resource crunch.
Any suggestions.
No Events found!


ble1
6 Operator
•
14.4K Posts
•
56.2K Points
0
January 6th, 2007 02:00
amediratta
2 Intern
•
2K Posts
0
January 6th, 2007 08:00
ble1
6 Operator
•
14.4K Posts
•
56.2K Points
0
January 7th, 2007 08:00
IKP1
198 Posts
0
January 9th, 2007 05:00
I am using AutoStart 5.1.2 with Exchange module.
I have configured NW 7.2 Build 172 and Exchange
module 4.1.
When we backup manually from either physical node, it
backs up very slow - 25 GB took 11 hours. When we run
server initiated backup for the virtual node, it
smoothly backs up one Storage group (5GB) but for the
second(25GB) - it starts off well but then backup
speed is OK easily around 10-12 MB/sec and backup
finishes in around 2 hours.
The problem is that the users get disconnected at
this time and get reconnected only when either the
backup is over or is terminated.
In a way, Exchange server is not usable during
backups. Both physical nodes are 4 GB Dell machines
with Windows 2003 & Exchange 2003 so there should not
be a resource crunch.
Any suggestions.
What levels of performance are you seeing for non Exchange backups of the two nodes except for the Exchange data?
You also did not detail what tape technology you are using, how it is connected to the NetWorker environment. Also, how is the shared data presented to the EAS domain - is this SAN attached storage or via shared SCSI and what type of disk array are you using..
Basically more detail required.....
IKP
ble1
6 Operator
•
14.4K Posts
•
56.2K Points
0
January 9th, 2007 09:00
amediratta
2 Intern
•
2K Posts
0
January 9th, 2007 09:00
We have a Gigabit network. Dell servers with 4 GB RAM each. One partition on each server of around 150 GB is used for Exchange database. They are mirrored to each other, there is no SAN storage.
Backups are on LAN and the library is SCSI. Backup server is Linux based.
The backups happen at a speed of around 8-10 MBps when server initiated and 3-4 MBps when client initiated. When server-initiated, the Exchange server gets disconnected from the users but the backups continue.
IKP1
198 Posts
0
January 11th, 2007 07:00
be full duplex forced on each end...
I agree with Hrvoje, have you verified that all the nics are forced to max speed and full duplex.
You also said SCSI storage - whats the interface and what disks are being shared.
From within AutoStart, how is the datasource configured ? Is this a shared disk environment or are you using one of the disk replication tools within EAS ?
Since this is a clustered application backup are you using NetWorker cluster client licences ?
IKP
Message was edited by:
IKP
Message was edited by:
IKP