duncan wrote:
>
> Every time someone starts /home/vpopmail/bin/vchkpw when checking their
> mail I am getting this error in syslog:
>
> Sep 8 13:08:59 masina kernel: portmap: RPC call returned error 111
> Sep 8 13:08:59 masina kernel: RPC: task of released request still
> queued!
> Sep 8 13:08:59 masina kernel: RPC: (task is on xprt_pending)
> Sep 8 13:08:59 masina kernel: lockd: failed to monitor 192.168.0.2
>
> The machine that is the pop server (masina) is using NFS to mount /home
> from a different machine (192.168.0.2). I know it's odd to be using
> /home/vpopmail over NFS like this, but I've got NFS working well. This
> is the only program that gives me NFS errors. Could someone suggest
> what the problem might be?
>
> Duncan
I think a good point to look at is the file locking code
in the int lock_it(char *lock_file) function in the
vpopmail.c file.
An interesting experiment, if you have the C skills, is
to replace the usage of lock_it with the new file
locking code from Eric Peters. The calls are in the
file_lock.c file and used in the bigdir.c code.
Eric had reported file corruption using the lock_it
code from vpopmail on the .dir-control file. So replaced
it with functions straight out of Richard Steven's
"Advanced Programming in the UNIX Environment"
So in terms of C code you could change to possibly fix
the problem, replace lock_it and unlock_it in vcdb.c
and vpopmail.c with the corresponding code from file_lock.
You can lift code straight from bigdir.c as an example.
Ken Jones
