All,
Apologies if this is a known issue, but I did dig around somewhat before
pestering everyone with a forum post...
I'm seeing a strange performance issue with a snv_133 NAS project.
Hardware is as follows:
Gigabyte GA-EP43-UD3L motherboard
Intel Celeron E3300 CPU
Disks are a mixture of sizes, but all Samsung Spinpoints
Drives are attached to the onboard ICH10 controller and 3 Lycom PE-103, dual
port PCIe Sata-II cards.
Basically, I'm seeing very poor i/o performance (c8 is the on-board, c11 is one
of the 3 PCIe cards):
r...@nas2:/# dd if=/dev/rdsk/c8t0d0s0 of=/dev/null bs=65536 count=16384
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 57.8887 s, 18.5 MB/s
r...@nas2:/#
And the associated iostat:
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
283.1 0.0 18120.1 0.0 0.0 0.5 0.0 1.8 0 50 c8t0d0
tty cpu
tin tout us sy wt id
0 54 0 81 0 19
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
282.6 0.0 18088.5 0.0 0.0 0.5 0.0 1.8 0 50 c8t0d0
r...@nas2:/# dd if=/dev/rdsk/c11t1d0s0 of=/dev/null bs=65536 count=16384
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 59.8362 s, 17.9 MB/s
r...@nas2:/#
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
275.4 0.0 17622.6 0.0 0.0 0.5 0.0 1.8 0 50 c11t1d0
tty cpu
tin tout us sy wt id
0 54 0 89 0 11
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
276.0 0.0 17663.1 0.0 0.0 0.5 0.0 1.8 0 50 c11t1d0
If I try a similar test from a Live Linux CD, I get 100+ MB/s read from both
controllers. And bizarrely, from the 2009.06 Opensolaris LiveCD, I get around
80MB/s read (more than enough for my purposes).
Digging further, I installed snv_111 on a spare disk and tried the test again
(80MB/s consistently), upgraded to snv_118 (same result). I couldn't seem to
get a stable upgrade beyond there, so I can't see where this issue actually
starts (or even if the opensolaris release is even a contributing factor).
Bottom line, it seems like disk (potentially sata only) performance has dropped
somewhere between snv_118 and snv_133.
Is anyone else seeing the same?
Can anyone help?
Scott
--
This message posted from opensolaris.org
_______________________________________________
perf-discuss mailing list
perf-discuss@opensolaris.org
