Anyone seeing any NFS lockups/weirdness with latest (ish) current??
Matt
matt at xtaz.co.uk
Thu Oct 23 10:09:31 PDT 2003
I have a typical NFS setup where I have two boxes both running -CURRENT
from around 24 hours ago. One box runs nfsd and rpcbind, and the other
nfsclient and mounts /usr/ports, /usr/src and /usr/obj read/write.
Since this latest cvsup I am finding that something is locking up. I am
currently trying to run a portupgrade on the nfs client which gets
randomly between 4-8% through downloading into the nfs mounted
/usr/ports/distfiles and then dies.
If I try a simple ls /usr/ports from another shell it hangs as well.
However an ls /usr/src works fine. So it is only the one mount that has
failed.
A ps of the processes shows this:
root 2071 0.0 0.2 772 560 p5 D+ 6:02PM 0:00.01 ls
usr/ports 0 1552 0 -1 0 nfsrcv
root 2054 0.0 0.5 2604 1272 p3 S+ 6:02PM 0:00.14
/usr/bin/fetch - 0 2023 0 -8 0 nfsaio
So they appear to be waiting for something nfs related.
I can not kill -9 those processes either. They do not respond. The only
way to make them exit is to umount -f /usr/ports where the processes
exit normally after a successfull unmount.
I have tried mounting using NFSv2, NFSv3, Over tcp and over udp, and
also tried IPv4 and IPv6. All give the same result.
Anyone else seeing weird behavior?
Matt.
More information about the freebsd-current
mailing list