Messages in this thread | | | Date | Thu, 10 Feb 2000 11:09:51 +0100 (CET) | Subject | Re: Again: Process stuck in D state while locking in NFS | From | Trond Myklebust <> |
| |
>>>>> " " == Miklos Szeredi <Miklos.Szeredi@eth.ericsson.se> writes:
>> > Which kernel are you running, and what is your setup?
> SuSE 6.2, kernel 2.2.14, Solaris 2.6 NFS server.
> I just noticed, that lockd wasn't running, so locking was > broken anyway.
This is a problem with the 2.2.14 locking. You need to manually start the 'lockd' daemon using the 'rpc.lockd' program provided in the nfs-utils package (ex-knfsd). If you don't already have this program, you can find a copy of nfs-utils at
ftp://ftp.linuxnfs.sourceforge.org/pub/nfs
My guess is that is the source of your problem. Please try this before you attempt any further debugging...
> It is not clear how to turn on RPC debugging. Hints?
echo '65535' >/proc/sys/sunrpc/rpc_debug
will turn on the full RPC debugging code. That will usually give syslogd a nasty turn though, since the volume of info is extremely large. I'd therefore initially start by doing this only when the hang occurs (after stopping all other NFS + locking activity from other processes).
> I can also start tcpdump but with a sufficiently good filter, > because this is not a reproduible thing, and I don't want to > fill up my hard disk. What sort of command line should I give > tcpdump?
Just start tcpdump when the hang occurs as per above:
tcpdump -vvv -s 256
Cheers, Trond
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |