Hello people,
Thanks again for your help with the pine-like "expunge" a couple weeks back.
I now have some questions regarding flock and fcntl.
My current arrangement is; Box B has an nfs share with users home
directories. Box A receives email with sendmail. Box A has the nfs share
from B mounted. As mail arrives, Box A processe a user procmail script (which
actually resides on B but the process runs on A) seperating out email for the
user into various folders. Box A and B are redhat 6.0 running 2.2.14 and
2.2.16 respectively. I am using mutt 1.2i, compiled from source by me all the
default options.
Later when I (the user) would run mutt from Box A, I would get mutt errors
that a mailbox "folder" (actually a file) could not be locked and that it was
read only. In my kernel messages on Box A I would get the following message
several times:
Box_A kernel: lockd: failed to monitor 192.168.1.5
I finally got some time to play with this today, and recompiled mutt with
"--enable-flock --disable-fcntl"... now I can delete email in folders!!! ...
however I still get the kernel messages in my logs.
So, I am asking...
- why would one choose flock or fcntl?
- could there be some negative repercussions from my current selection?
- any other insights on how to correct/stop these nfs locking errors?
thanks,
donfede
