NWK_LG ドメインが高いため、 CPU の utlization が高くなっています
環境
問題
- CPU 利用率が高い
例:
cluster::*> run -node node1 sysstat 1
CPU NFS CIFS HTTP Net kB/s Disk kB/s Tape kB/s Cache
in out read write read write age
53% 5324 48229 0 38327 6833 127216 390644 0 0 6s
64% 6208 38916 0 64143 10034 419764 1445352 0 0 6s
50% 26610 51969 0 73790 8921 119760 382280 0 0 6s
53% 29072 57445 0 125694 10797 106164 408976 0 0 6s
89% 15387 45833 0 120733 8356 176208 521896 0 0 10s
99% 15003 48234 0 131641 7639 190944 511900 0 0 2
99% 12012 64723 0 105238 16576 373572 512840 0 0 0s
99% 15154 35868 0 148367 13014 255076 405512 0 0 2
99% 12834 34674 0 115071 6865 215028 550032 0 0 2
99% 12465 32334 0 133452 17632 414064 679288 0 0 2
99% 13036 37854 0 126981 8369 204972 402760 0 0 3
CPU NFS CIFS HTTP Net kB/s Disk kB/s Tape kB/s Cache
in out read write read write age
99% 13897 33801 0 141796 3869 104280 335868 0 0 2
99% 12403 30245 0 87923 9940 319380 821176 0 0 2s
99% 10065 29574 0 172745 8461 158264 417404 0 0 2s
99% 11967 33489 0 154454 9136 206668 564908 0 0 3
99% 11428 34399 0 81521 3313 136904 391944 0 0 2
- 高い NWK_LG ドメインで CPU 使用率が高くなっています。
cluster::> node run -node <node_name>
node> priv set diag
node*> sysstat -M
例:
fragments received
の出力からの膨大な量"netstat -s"
。
例:
> ip:
> 645136792449 total packets received
> 154113333490 fragments received <<< Here>>>
> 7271533072 packets reassembled ok
> 464498769419 packets sent from this host
> 801059272 output datagrams fragmented
> 13545054029 fragments created
> icmp:
> Output histogram:
> echo reply: 19561080
> destination unreachable: 273761
> Input histogram:
> echo reply: 5331332
> destination unreachable: 708
> echo: 19561716
> time exceeded: 8556
- UDP 経由の NFS :