HI,boss
I have read "Open Storage with the Solaris ZFS File System and
COMSTAR iSCSI". As you mentioned in the document,wo should use
OpenSolaris 0906. But the performance is so bad. Then we update the
image to b118,The performance is better.And OpenSolaris 08.11 's is
better than 0906.
Now I have a problem. I use two intel e1000g nic card(chip 82563)
to form an aggregation. The bandwidth is balance.
The result is below.
LINK PORT IPACKETS RBYTES OPACKETS OBYTES IPKTDIST OPKTDIST
aggr1 -- 64285 97567422 38762 2569640 -- --
-- e1000g0 32477 49291258 19108 1261228 50.5 49.3
-- e1000g1 31808 48276164 19654 1308412 49.5 50.7
aggr1 -- 61136 92780290 36960 2450274 -- --
-- e1000g0 31537 47861814 18514 1231464 51.6 50.1
-- e1000g1 29599 44918476 18446 1218810 48.4 49.9
aggr1 -- 67188 101964782 40693 2705000 -- --
-- e1000g0 33875 51408046 19893 1322432 50.4 48.9
-- e1000g1 33313 50556736 20800 1382568 49.6 51.1
aggr1 -- 66847 101455658 40405 2680660 -- --
-- e1000g0 33884 51427158 19782 1309300 50.7 49.0
-- e1000g1 32963 50028500 20623 1371360 49.3 51.0
aggr1 -- 67062 101776254 40579 2695138 -- --
-- e1000g0 33329 50581988 19733 1308412 49.7 48.6
-- e1000g1 33733 51194266 20846 1386726 50.3 51.4
aggr1 -- 67977 103167918 40999 2716026 -- --
-- e1000g0 34619 52542548 20093 1334764 50.9 49.0
-- e1000g1 33358 50625370 20906 1381262 49.1 51.0
aggr1 -- 57673 87522524 35156 2341246 -- --
-- e1000g0 29900 45380660 17481 1164942 51.8 49.7
-- e1000g1 27773 42141864 17675 1176304 48.2 50.3
But the bandwidth is just two 40MB/s.
The mpstat result is :
r...@opensolaris:~# mpstat 1
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 25 0 12 1117 106 21202 24 1697 3304 580 169 0 35 0 65
1 22 0 16 933 12 22236 42 2054 2521 826 251 0 25 0 75
2 32 0 37 569 13 12855 10 1869 2538 747 282 0 25 0 75
3 22 0 94 5396 4718 12952 65 1799 2163 741 210 0 25 0 75
4 42 0 147 92 12 5790 4 2031 2031 938 400 1 13 0 85
5 36 0 107 107 11 6323 5 2026 2143 933 485 1 14 0 85
6 44 0 143 101 13 5930 5 1953 2037 913 361 1 14 0 85
7 43 0 165 8624 8494 5548 7 1749 2077 742 445 1 25 0 74
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 304 102 21 0 3 2 0 2 0 0 0 100
1 0 0 0 22 2 48 0 0 1 0 0 0 1 0 99
2 0 0 0 57 23 195 0 1 1 0 30 0 2 0 98
3 0 0 0 53 25 78 0 3 0 0 58 0 0 0 100
4 28 0 0 25 4 44 0 4 1 0 306 0 1 0 99
5 0 0 0 11 1 12 0 1 1 0 0 0 0 0 100
6 0 0 0 81 34 107 0 3 1 0 0 0 1 0 99
7 0 0 0 160 137 14 0 1 1 0 0 0 0 0 100
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 300 100 41 0 5 2 0 0 0 0 0 100
1 0 0 0 14 1 54 0 1 1 0 0 0 0 0 100
2 0 0 0 37 15 184 0 2 3 0 2 0 2 0 98
3 0 0 0 18 7 55 0 3 5 0 61 0 1 0 99
4 5 0 0 22 2 28 0 3 1 0 304 0 0 0 100
5 0 0 0 73 32 74 0 3 2 0 0 0 0 0 100
6 0 0 0 36 14 106 0 3 0 0 0 0 0 0 100
7 0 0 0 175 151 28 0 3 1 0 0 0 0 0 100
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 601 100 18835 0 952 6362 32 0 0 43 0 57
1 0 0 0 579 14 23012 0 1462 1982 30 0 0 26 0 74
2 0 0 0 338 10 21717 0 1425 3016 27 2 0 37 0 63
3 0 0 0 8918 8377 21569 53 1301 2772 33 61 0 34 0 66
4 0 0 0 71 13 4511 0 1147 1631 41 300 0 11 0 89
5 0 0 0 65 10 7139 0 1344 1434 33 0 0 14 0 86
6 0 0 0 57 6 5504 0 1172 1741 32 0 0 13 0 87
7 0 0 0 7893 7817 7579 2 1520 2500 37 0 0 24 0 76
My hard raid is raid 5 with 10 disks,and has been set write back
mode.The aggregation si set policy as L4.The aggregation without
switch's LACP support.I use two windows 2003 server as client.And the
test tools is IOmeter.The result came with 8 workers in a iometer.
What the problem is ?
Could you help me?
Thank you very much !
yours,dai
--
This message posted from opensolaris.org
_______________________________________________
networking-discuss mailing list
[email protected]
