cancel
Showing results for 
Search instead for 
Did you mean: 

Poor archiving performance

AndresMunoz
Level 5
Partner Accredited

I'm currently experiencing VERY poor archiving performance on a newly built 10.0.4 environment.

Environment Details

  • VMware ESX 5.1.
  • All virtual servers are using vmxnet3 NIC
  • 200 Online Outlook 2010 clients
  • 42 ActiveSync users (approximately 80 Apple devices)
  • Average item size 70KB

EV Server

  • Windows Server 2008 R2 SP1
  • Enterprise Vault 10.0.4
  • Outlook 2007 SP3

SQL Server

  • Windows Server 2008 R2 SP1
  • MS SQL 2012 Standard

Exchange Server

  • Windows Server 2008 R2 SP1
  • Exchange Sever 2010 SP2 CU 4 (cannot install latest due to compatibility with other customer systems)
  • 2 Servers: CAS+MBX+HT and CAS+MBX

I have built the system and applied all recommendations on the EV 10 performance guide and VMware whitepaper, yet I'm getting approximately 800 MB hour archival rate.

I'm certain the EV server configuration is not the problem, I've deployed similar configuration on other sites, and also on my lab, and I get between 10-12 GB/hour of of a single server.

I believe the problem is exchange. I did a two test mailbox moves of 300MB and 870MB. they moved in 30min and 1hour 7 minutes respectively. on other environments we normally transfer between 15-20 GB hour

I've run out of ideas with regards to tweaking the exchange performance. All exchange performance counters indicate the systems are doing well without obvious bottlenecks. CPU usage increases to 90% during archiving run.

Any help would be greatly appreciate. I can provide more information if needed as I'm aware the. suggestions, pointers, past experience would be welcomed.

1 ACCEPTED SOLUTION

Accepted Solutions

JesusWept3
Level 6
Partner Accredited Certified
Do you use a CAS array at all? Have seen it when load balances can slow things down when they're not optimized Also maybe worth checking the network and NIC configuration on the exchange servers, sometimes I've seen it where the backup nics are used Also make sure silly things like the TCP Chimney/Offload is disabled on the EV server and the exchange server May also be worth looking at the disk queues on the exchange side of things too
https://www.linkedin.com/in/alex-allen-turl-07370146

View solution in original post

2 REPLIES 2

JesusWept3
Level 6
Partner Accredited Certified
Do you use a CAS array at all? Have seen it when load balances can slow things down when they're not optimized Also maybe worth checking the network and NIC configuration on the exchange servers, sometimes I've seen it where the backup nics are used Also make sure silly things like the TCP Chimney/Offload is disabled on the EV server and the exchange server May also be worth looking at the disk queues on the exchange side of things too
https://www.linkedin.com/in/alex-allen-turl-07370146

Rob_Wilcox1
Level 6
Partner

Is there any more help needed on this issue?

Working for cloudficient.com