Unsolved
This post is more than 5 years old
2 Posts
0
822
March 22nd, 2010 04:00
Slow replication performance in MirrorView/A over iSCSI
We have two CX3-20c in Production and DR site. The replication between two sites is using MirrorView/A over iSCSI.
The link between two sites is 100MB/s and the latency is very low i.e. 1ms.
We would like to replicate 2 x 500GB from Production to DR but found that it is only replicating 100GB over last thee weeks. The sync rate is medium. We have checked with our Network team that the bandwidth is not over-utlizated. It seems that the CLARiiON cannot use up the bandwidth.
Do you guys have any idea on it?
No Events found!
kelleg
6 Operator
•
4.5K Posts
0
March 23rd, 2010 15:00
Do you have hosts configured on the Mirrorview ports?
Have you tried changing the sync rate to high?
Initial sync can take a long time - although 3 weeks seems excessive for 500GB.
Are you seeing any error messages in the SP Event Log?
Are you configured for jumbo frames?
In almost cases that I've seen the issue have been the network - somewhere
glen
chung_chow
2 Posts
0
March 24th, 2010 06:00
The hosts are connected to other SP port which is fabric while the MirrorView port is iSCSI.
The replication is still slow even though the sync rate is high.
No error in the SP evenet log as everything is normal during the replication.
I haven't configured the iSCSI port for jumbo frames. Does it faster after configured for jumbo frames.
kelleg
6 Operator
•
4.5K Posts
0
March 24th, 2010 10:00
Jumbo frames can help, but all devices in the network must be capably of supporting jumbo frames.
I would still recommend looking at the network - if there are no errors in the logs, that indicates that the connection is not going up and down. See if you can get your network people to look at the paths that the data is taking.
You also might want to look at the target array to ensure that the destination LUNs have the capacity to handle the IO load.
glen