Hi,
We have a storage server (HP DL360G5 + MSA20 (12 disks in RAID 6) on a SmartArray6400).
10 directories are exported through nfs to 10 clients (rsize=32768,wsize=32768,soft,intr,nosuid,proto=udp,vers=3).
The server is apparently not doing much but... we have very high waiting IOs.
dstat show very little activity, but high 'wai'...
# dstat
----total-cpu-usage---- -dsk/total- -net/total- ---paging-- ---system--
usr sys idl wai hiq siq| read writ| recv send| in out | int csw
0 0 88 12 0 0| 413k 98k| 0 0 | 0 0 | 188 132
0 1 46 53 0 0| 716k 48k| 19k 420k| 0 0 |1345 476
0 1 49 50 0 1| 492k 32k| 12k 181k| 0 0 |1269 482
0 1 63 37 0 0| 316k 159k| 58k 278k| 0 0 |1789 1562
0 0 74 26 0 0| 84k 512k|1937B 6680B| 0 0 |1200 106
0 1 44 55 0 1| 612k 80k| 14k 221k| 0 0 |1378 538
1 1 52 47 0 0| 628k 0 | 17k 318k| 0 0 |1327 520
0 1 50 49 0 0| 484k 60k| 14k 178k| 0 0 |1303 494
0 0 87 13 0 0| 124k 0 |7745B 116k| 0 0 |1083 139
0 1 59 41 0 0| 316k 60k|4828B 67k| 0 0 |1179 346
top shows that one nfsd is usualy in state 'D' (waiting).
# top -i (sorted by cpu usage)
top - 18:11:28 up 207 days, 7:13, 2 users, load average: 0.99, 1.07, 1.00
Tasks: 124 total, 1 running, 123 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.2%sy, 0.0%ni, 54.3%id, 45.3%wa, 0.2%hi, 0.0%si, 0.0%st
Mem: 3089252k total, 3068112k used, 21140k free, 928468k buffers
Swap: 2008116k total, 164k used, 2007952k free, 293716k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+
COMMAND
16571 root 15 0 12708 1076 788 R 1 0.0 0:00.02
top
2580 root 15 0 0 0 0 D 0 0.0 2:36.70 nfsd
# cat /proc/net/rpc/nfsd
rc 8872 34768207 38630969
fh 142 0 0 0 0
io 2432226534 884662242
th 32 394 4851.311 2437.416 370.949 238.432 542.241 4.942 2.239 1.000 0.427 0.541
ra 64 3876274 5025 3724 2551 2030 2036 1506 1607 1219 1154 1136249
net 73410453 73261524 0 0
rpc 73408119 0 0 0 0
proc2 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
proc3 22 33 9503937 1315066 11670859 7139862 0 5033349 28129122 3729031 0 0 0 487614 0 1116215 0 0 2054329 21225 66 0 2351744
proc4 2 0 0
proc4ops 40 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Do you think nfs is the problem here?
If so, is there something wrong with our config?
Is it too much to have 10 dir x 10 clients, even if there is almost no traffic?
Thx,
JD