Hi again,
Further to my previous posts, I am seeing issues over time with zfs performance.
If I leave the system running for 2+ weeks and do the following benchmarks I get:
Rebooting and doing nothing I am seeing:
Which is 2x quicker.
I am running 9.0-RELEASE on an AMD Athlon II x4 630 with 10G of memory. The array is a RAIDZ of 4 x Samsung 2TB disks that are aligned with ashift=12.
Any ideas? (I can see a long debug cycle as I wait for this to slow again...).
Further to my previous posts, I am seeing issues over time with zfs performance.
If I leave the system running for 2+ weeks and do the following benchmarks I get:
Code:
[andy@MAINSERVER ~]$ sudo bonnie++ -d /storage/temp/dir -u 0:0 -s 20g
...
Version 1.96 ------Sequential Output------ --Sequential Input- --Random-
Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
MAINSERVER 20G 150 99 99443 26 54030 13 340 91 140299 14 105.4 7
Latency 93065us 6100ms 2606ms 630ms 568ms 1210ms
Version 1.96 ------Sequential Create------ --------Random Create--------
MAINSERVER -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 24099 91 +++++ +++ 16864 67 23505 87 +++++ +++ 25727 96
Latency 10047us 116us 8712us 18544us 40us 95us
Rebooting and doing nothing I am seeing:
Code:
[andy@MAINSERVER ~]$ sudo bonnie++ -d /storage/temp/dir -u 0:0 -s 20g
...
Version 1.96 ------Sequential Output------ --Sequential Input- --Random-
Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
MAINSERVER 20G 151 99 223546 58 105195 25 366 97 285066 28 104.2 6
Latency 152ms 2482ms 2432ms 121ms 462ms 908ms
Version 1.96 ------Sequential Create------ --------Random Create--------
MAINSERVER -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 23602 91 +++++ +++ 21375 90 20619 85 +++++ +++ 19788 74
Latency 12456us 125us 177us 30806us 77us 2527us
Which is 2x quicker.
I am running 9.0-RELEASE on an AMD Athlon II x4 630 with 10G of memory. The array is a RAIDZ of 4 x Samsung 2TB disks that are aligned with ashift=12.
Any ideas? (I can see a long debug cycle as I wait for this to slow again...).