No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

Storage write performance deteriorates due to inaccurate number of disks in a mirror group of RAID 10 for the S5300

Publication Date:  2012-07-20 Views:  54 Downloads:  0
Issue Description
On a deployment site, fourteen 300 GB FC disks are connected to a Linux server by using the FC host port. The 14 disks form a RAID 10 group and are divided into two 1 TB LUNs. The two LUNs are mapped to the Linux server. The Linux server is logged in to and the read/write performance of the disks mapped by the S5300 is tested. It takes five minutes to create a 1 GB file. The write performance is poor.
Alarm Information
None
Handling Process

Step 1  Log in to the Linux server and uninstall LUNs.
Step 2  Log in to the S5300 management interface and delete LUNs and the RAID group.
Step 3  Create a new RAID 10 group, set the number of disks in the mirror group to two, and select all 14 disks.
Step 4  Create two 1 TB LUNs and map them to the server.
Step 5  Log in to the Linux server, rescan disks, identify the two LUNs, partition and format the two LUNs respectively, and mount LUNs.
Step 6  Test the read/write performance of the disks. It takes 30 seconds to create a 5 GB file. The write performance obviously improves.
Root Cause
When 14 disks form one mirror group of a RAID 10 group, the capacity of the RAID group is equal to one disk; therefore, the write performance is poor. Usually, the number of disks in a mirror group is two. The capacity of the RAID group is 50% of the raw capacity and its capacity is between RAID 0 and RAID 5.
Suggestions
During the deployment, install and configure storage devices strictly according to the deployment guide.

END