- Sep 20, 2014
- 4
- 0
- 4,510
I'm running an IO test using IOMeter (downloaded from Sourceforge) on a ML380 G5 with (4) 10k SAS drives in a RAID 5 configuration. The issue the server is experiencing is the test is coming back with results of about 160 iOPS/s. I have (2) other servers, same make and model, same hardware configuration, in production, and those are returning results of roughly 3000 iOPS/s. There is no hardware or configuration difference between the production servers and the non-production server.
The test has been ran multiple times on all (3) servers, each time resulting in the expected output for the production servers, but not the non-production server. My non-production server should have the same (or relatively close to) results as my (2) production servers.
I've waited for the RAID to fully build before running the test. I'm beginning to believe the hard drive(s) may be the problem. If at least one drive is not performing up to standards, then that would cause the issue the server is experiencing for the I/O test.
However, I would like to hear some other thoughts and ideas on what the other possible causes could be that I have not considered.
Thanks.
The test has been ran multiple times on all (3) servers, each time resulting in the expected output for the production servers, but not the non-production server. My non-production server should have the same (or relatively close to) results as my (2) production servers.
I've waited for the RAID to fully build before running the test. I'm beginning to believe the hard drive(s) may be the problem. If at least one drive is not performing up to standards, then that would cause the issue the server is experiencing for the I/O test.
However, I would like to hear some other thoughts and ideas on what the other possible causes could be that I have not considered.
Thanks.
Facebook
Twitter
Instagram