Subject: hanging nfs server?
To: None <current-users@netbsd.org>
From: matt <deberg@xennahtron.com>
List: current-users
Date: 10/22/2000 19:37:42
i've got a netbsd/i386 box running 1.5F and exporting a bunch of NFS
volumes. it will reliably hang after moderate/heavy use after a few
hours, such as a diskless netbsd client running make build, all over
nfs.
on the server, i've got this:
[chamomile!deberg] ps alx | grep nfsd
ps: warning: /var/run/dev.db: No such file or directory
0 236 1 29 2 0 68 536 netcon Is ?? 0:00.01 nfsd: master
0 252 236 0 -5 0 48 428 fsync_ra DL ?? 3:31.66 nfsd: server
0 253 236 1 -2 0 48 428 vnlock DL ?? 0:06.90 nfsd: server
0 254 236 0 -2 0 48 428 vnlock DL ?? 0:01.01 nfsd: server
0 255 236 0 -2 0 48 428 vnlock DL ?? 0:00.41 nfsd: server
19890 8307 7587 0 28 0 540 1152 - RV ?? 0:00.00 grep nfsd (tcs
and on a client that has a hung process doing ls on the mountpoint:
nonantum-road# ps alx | grep ls
0 26682 26674 0 -1 0 352 88 nfsrcvlk D p4 0:00.01 ls /home
0 26699 26674 0 28 0 548 152 - RV p4 0:00.00 grep ls (csh)
getting a stack trace of an nfsd on the server is hard, because it's the
important machine, but it looks something like this:
bpendtsleep
lockmgr
genfs_lock
vn_lock
vget
ufs_ihashget
ffs_vget
ufs_fhtovp
ffs_fhtovp
nfsrv_fhtovp
nfsrv_commit
nfssvc_nfsd
sys_nfssvc
any idea what's going on? i'll leave the machine in this state for
awhile in case there's more useful information i can get.
matt