From: tsh@mrc-lmb.cam.ac.uk
Date: Fri Feb 21 2003 - 07:29:22 EST
Hi all,
TU64 v5.1 PK4 on an ES40 4cpu 8Gb mem.
I've noticed high cpu use by the kernel process over the last couple
of days. Usage has been between 40 and 90% cpu more-or-less continuously.
vmstat and nfsstat show nothing abnormal, and a trawl through the archives
suggested doing 'ps -p 0 -m -o wchan,state,time'.
This shows something interesting. If I repeat this command continuously,
the TIME value for the tcpfstm thread clocks up at about the same rate as
the kernel cpu usage. i.e. this thread seems to account for all the
high kernel use I'm seeing. What does tcpfstm do?
It doesnt always appear in the output from the ps command, but when
it does, its always in state 'S <' .
Any ideas anyone?
Cheers,
Terry.
WCHAN S TIME
* R < 25-13:15:19
- R N 0:00.00
malloc_ S < 1:01.67
24100d8 U < 0:02.96
ubc_pur I < 5:37.48
Tswpin I < 0:00.00
2410398 U < 0:00.00
2410418 U < 0:00.00
8f8cf0 I < 0:03.88
8faba8 S < 0:01.16
8fabe8 I < 0:00.03
- R N 0:00.00
- R N 0:00.00
- R N 0:00.00
c46080 U < 0:00.51
camwork S < 10:13:13
camwork I < 0:00.00
a91e58 U < 0:00.41
a8eaf0 U < 0:19.52
a8e2b0 U < 0:28.67
ss_tmo S < 0:08.95
ss_tmo S < 0:09.44
ss_tmo S < 0:08.84
ss_tmo S < 0:09.04
ss_tmo S < 0:08.82
ss_tmo S < 0:08.97
ss_tmo S < 0:09.21
ss_tmo S < 0:07.24
isp_rq I < 2:24.93
isp_abo I < 0:00.00
isp_fm I < 0:00.00
ss_tmo S < 0:06.64
ss_tmo S < 0:02.41
6fe170 U < 0:00.00
altrecv S < 22:19:58
altxmit S < 17:14:56
ss_tmo S < 0:06.23
8fb840 U < 0:00.00
ee1c1500 U < 0:00.00
ss_tmo S < 0:02.61
ee1c0540 U < 0:00.00
ss_tmo S < 0:02.31
edf75880 U < 0:00.00
ede02000 U < 0:00.01
8faf90 U < 0:00.00
ede12128 U < 0:00.05
c46248 U < 0:00.00
netisr S < 04:53:56
netisr S < 04:52:53
netisr S < 04:48:59
netisr S < 04:52:59
netisr S < 04:50:07
tcpfstm S < 7-03:54:19
tcpsltm S < 0:38.18
8a1df9f0 S < 2:32.39
netrain I < 0:00.00
netrain S < 0:01.11
netrain I < 0:00.00
8fb280 U < 0:01.16
66e3f0 U < 0:00.02
6610c0 U < 0:29.75
661300 U < 0:00.00
ed87a680 U < 0:00.00
ed87a640 U < 0:00.22
ed87a600 U < 0:00.29
ed87a5c0 U < 0:00.28
ed87a580 U < 0:00.41
8a2479f0 S < 0:01.58
8f9738 S < 0:00.48
96c3b0 U < 0:00.00
96c3b0 U < 0:00.00
96c3b0 U < 0:00.00
96c3b0 U < 0:00.00
atm_arp I < 0:00.00
8a27f9f0 S < 0:12.56
atmifmp I < 0:00.00
959858 U < 0:00.00
95a2b8 U < 0:00.00
25b5c0 U < 0:00.48
1383140 U < 10-05:30:47
8f7c1c U < 0:00.00
a8cc20 U < 0:00.00
a8cc20 U < 0:00.00
d7ca0a30 U < 5:32.30
d7ca09d0 U < 0:00.00
d7ca0970 U < 0:03.42
d7ca0910 U < 0:02.44
d7ca08b0 U < 0:24.70
nfs_tcp I 0:00.00
nfs_tcp I 0:00.00
nfs_tcp I 0:00.00
nfs_tcp I 0:00.00
nfs_tcp I 0:00.00
nfs_tcp I 0:00.00
nfs_tcp I 0:00.00
nfs_tcp I 0:00.00
nfs_udp S 15:06:48
1e52b40 U 15:08:04
bb1e7164 U 15:08:43
nfs_udp S 15:08:57
nfs_udp S 15:07:25
nfs_udp I 15:09:03
nfs_udp I 15:06:47
nfs_udp I 15:07:16
nfsiod_ I 2:11.38
nfsiod_ I 2:33.49
nfsiod_ I 3:40.23
nfsiod_ I 2:19.46
nfsiod_ I 2:28.79
nfsiod_ I 1:40.13
nfsiod_ I 3:08.99
a96eb0 U < 0:00.00
#
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:49:08 EDT