While not strictly a ZFS issue as such I thought I'd post here as this and the
storage forums are my best bet in terms of getting some help.
I have a machine that I recently set up with b130, b131 and b132. With each
build I have been playing around with ZFS raidz2 and mirroring to do a little
performance testing. This is a 6 SATA port ICH10 Intel motherboard, running in
AHCI mode. OS is on a USB flash. Suffice to say I have noticed that 1
particular drive out of 6 seems to have very high asvc_t practically all the
time. This is an excerpt from 'iostat -xnM c6t2d0 2';
[i] r/s w/s Mr/s Mw/s wait actv wsvc_t asvc_t %w %b device
70.5 502.0 0.0 4.1 0.0 1.3 0.0 2.2 0 54 c6t2d0
50.5 137.5 0.0 3.0 0.0 0.7 0.0 3.9 0 47 c6t2d0
71.0 163.5 0.0 4.8 0.0 0.8 0.0 3.4 0 61 c6t2d0
13.5 29.5 0.0 1.0 0.0 2.6 0.0 61.4 0 88 c6t2d0
1.0 0.5 0.0 0.0 0.0 3.6 0.0 2406.2 0 100 c6t2d0
1.0 1.0 0.0 0.0 0.0 4.0 0.0 1993.4 0 100 c6t2d0
1.0 1.5 0.0 0.0 0.0 4.0 0.0 1593.8 0 100 c6t2d0
2.0 3.0 0.0 0.1 0.0 4.0 0.0 791.6 0 100 c6t2d0
1.0 2.0 0.0 0.1 0.0 4.0 0.0 1320.3 0 100 c6t2d0
1.0 5.0 0.0 0.3 0.0 3.6 0.0 595.1 0 100 c6t2d0
[/i]
and here is the drive shown with the other in the raidz2 pool
extended device statistics
r/s w/s Mr/s Mw/s wait actv wsvc_t asvc_t %w %b device
0.0 1.5 0.0 0.0 0.0 0.0 0.0 0.5 0 0 c6t0d0
0.0 1.5 0.0 0.0 0.0 0.0 0.0 0.3 0 0 c6t1d0
1.0 1.0 0.0 0.0 0.0 4.0 0.0 1994.8 0 100 c6t2d0
1.0 1.5 0.0 0.0 0.0 0.0 0.0 5.2 0 1 c6t3d0
1.0 1.5 0.0 0.0 0.0 0.0 0.0 6.9 0 1 c6t4d0
1.0 1.5 0.0 0.0 0.0 0.0 0.0 10.1 0 2 c6t5d0
extended device statistics
r/s w/s Mr/s Mw/s wait actv wsvc_t asvc_t %w %b device
1.0 5.5 0.0 0.2 0.0 0.0 0.0 1.6 0 1 c6t0d0
1.0 5.5 0.0 0.2 0.0 0.0 0.0 1.5 0 1 c6t1d0
2.0 3.5 0.0 0.1 0.0 4.0 0.0 721.8 0 100 c6t2d0
1.0 5.5 0.0 0.2 0.0 0.0 0.0 1.9 0 1 c6t3d0
1.0 5.5 0.0 0.2 0.0 0.0 0.0 1.6 0 1 c6t4d0
2.0 5.5 0.0 0.2 0.0 0.0 0.0 3.1 0 2 c6t5d0
extended device statistics
r/s w/s Mr/s Mw/s wait actv wsvc_t asvc_t %w %b device
0.0 3.5 0.0 0.1 0.0 0.0 0.0 0.4 0 0 c6t0d0
0.0 3.5 0.0 0.1 0.0 0.0 0.0 1.8 0 0 c6t1d0
1.0 2.0 0.0 0.1 0.0 4.0 0.0 1327.1 0 100 c6t2d0
1.0 3.5 0.0 0.1 0.0 0.0 0.0 4.9 0 1 c6t3d0
1.0 3.5 0.0 0.1 0.0 0.0 0.0 3.9 0 1 c6t4d0
1.0 3.5 0.0 0.1 0.0 0.0 0.0 2.0 0 1 c6t5d0
I have seen asvc_t as high as 20000.
There do not appear to be any errors hardware wise as 'iostat -e' shows
[i] ---- errors ---
device s/w h/w trn tot
sd0 0 0 0 0
sd2 0 0 0 0
sd3 0 0 0 0
sd4 0 0 0 0
sd5 0 0 0 0
sd6 0 0 0 0
sd7 0 0 0 0
[/i]
'zpool iostat -v 2' pauses for anywhere between 3 and 10 seconds before it
prints the stats for that particular drive in the pool;
capacity operations bandwidth
pool alloc free read write read write
------------------ ----- ----- ----- ----- ----- -----
data 185G 5.26T 3 115 8.96K 2.48M
raidz2 185G 5.26T 3 115 8.96K 2.48M
c6t0d0 - - 2 26 2.70K 643K
c6t1d0 - - 2 26 2.49K 643K
* INSERT ~10 SECOND PAUSE*
c6t2d0 - - 2 24 2.81K 643K
c6t3d0 - - 2 26 2.75K 643K
c6t4d0 - - 2 26 2.45K 643K
c6t5d0 - - 2 26 2.71K 643K
------------------ ----- ----- ----- ----- ----- -----
rpool 3.50G 3.94G 0 0 9.99K 1010
c5t0d0s0 3.50G 3.94G 0 0 9.99K 1010
------------------ ----- ----- ----- ----- ----- -----
swpool 102K 3.69G 0 0 19 0
/dev/rdsk/c7t0d0 102K 3.69G 0 0 19 0
------------------ ----- ----- ----- ----- ----- -----
I have booted up a linux rescue CD that has S.M.A.R.T support (system rescue
CD) and performed the 'long' test on each drive. All drives pass the 'test'.
There also appears to be no system errors with the drives under linux either.
Can anyone shed any light on this issue, or suggest what I could try next ? I
am sort of discounting hardware problems given that I do not see errors from
the live linux CD. Maybe I should install linux and see if the problem persists
?
Cheers.
--
This message posted from opensolaris.org
_______________________________________________
storage-discuss mailing list
storage-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
