Unsolved
This post is more than 5 years old
8 Posts
0
1129
January 13th, 2017 00:00
ScaleIO 2.0.1.2 on XenServer 6.5 with RF Cache
Guys, we have an environment with 3 nodes. Each node have 8x sata 6TB and 2x ssd 200GB for rfcache.
After insert and enable disks for rfcache xenserver is showing log below:
messages:Jan 10 01:47:04 p01-c01-h03 kernel: [ 2730.123284] 8:112: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:06 p01-c01-h03 kernel: [ 2733.083229] 8:80: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:07 p01-c01-h03 kernel: [ 2733.494498] 8:176: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:08 p01-c01-h03 kernel: [ 2734.380540] 8:112: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:08 p01-c01-h03 kernel: [ 2735.026868] 8:128: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:09 p01-c01-h03 kernel: [ 2735.239249] 8:160: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:09 p01-c01-h03 kernel: [ 2735.616695] 8:64: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:09 p01-c01-h03 kernel: [ 2735.896629] 8:96: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:11 p01-c01-h03 kernel: [ 2737.124595] 8:128: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:11 p01-c01-h03 kernel: [ 2737.456669] 8:144: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:11 p01-c01-h03 kernel: [ 2737.719336] 8:80: request aux data allocation failed, iosched may be disturbed
messages:Jan 10 01:47:12 p01-c01-h03 kernel: [ 2738.587153] 8:128: request aux data allocation failed, iosched may be disturbed
I tried with one and two disk in rfcache, same log result. Tested with io scheduler cfq (default) and noop
Any help?
Regards.
Marcelo Lima
pawelw1
306 Posts
0
January 13th, 2017 06:00
Hi Marcelo,
I couldn't find such an error message in our kbase or tickets. I only managed to find this URL:
https://github.com/datto/dattobd/issues/6
Seems the problem might be indeed with the scheduler. Can you try to change it to "deadline" (capture performance results before changing it and after the change) and see if it goes away?
Cheers,
Pawel
marcelolotic
8 Posts
0
January 16th, 2017 23:00
Hi Pawel.
I changed scheduler to deadline, no errors in hypervisor
Now Read Hit Cache is 58%
About performance I use bonnie++ but the result is very confuse, can you recommend another tool to test IOPs (iometer, iozone…)?
Thank you for help
--
Marcelo Lima
pawelw1
306 Posts
0
January 23rd, 2017 06:00
Hi Marcelo,
Never used bonnie++, so can't really comment on it - usually we run 'fio' or 'iometer'.
Many thanks,
Pawel