Original Question:
------------------
-> According to the automount man page ...
->
-> automount must not be terminated with the SIGKILL signal
-> (kill -9). Without an opportunity to unmount itself, the
-> automount mount points will appear to the kernel to belong
-> to a non-responding NFS server. The recommended way to ter-
-> minate automount services is to send a SIGTERM (kill -15)
-> signal to the daemon.
->
-> Unfortunately, I did just this, i.e., kill automount with -9.
-> I can't remove the mount point or even ls the directory it's in.
-> Restarting the automounter didn't help. Any ideas on how I can get
-> out of this?
->
-> By the way, this problem arose trying to fix another problem. The
-> machine in question had been automounting /var/spool/mail from another
-> machine. However, mail would occassionally fail on the machine doing
-> the automounting with messages like
->
-> Oct 30 18:46:05 seneca sendmail[6086]: AA06086: SYSERR: net hang reading
-> from cayuga.cs.jmu.edu: Connection timed out during collect with
-> cayuga.cs.jmu.edu
->
-> Oct 30 18:46:04 cayuga.cs.jmu.edu sendmail[12807]: NOQUEUE: SYSERR:
-> net hang reading from seneca: Connection timed out
->
The consensus is that both (kill -15) and (kill -9) can bring about this
problem and that the easiest solution is to just reboot. However, even
this may not be sufficient. The first response (which came moments after
I posted) was from Richard Feuerriegel, and contains all of the possible
solutions I received.
Richard writes
> If my machine dies while automount is still active, usually the fsck on
> reboot will take care of the problem directories. Sometimes it takes a
> manual run though fsck to clear the bad dirs. Other times, I have to
> reboot in single user mode and remove the directories.
------------------------------------------------------------------------------
With regard to the sendmail error, Robert Montjoy writes
> The "net hung" reading from bla bla bla is caused by not having a carriage
> return or is it a line feed at the end of that particular mail message.
> If you are having the problem often make sure your users have a blank line
> at the end of the mail message(.signature file).
In addition, Steve Swaney writes
> We had the same problem. I had nothing to do with mail. The "finger"
> command was also "broke". In fact anything that did a username lookup
> was broken except "ypcat passwd". The problem was that one of the
> passwd entries in the NIS passwd map had a "," instead of a ":". When
> sendmail (or finger, or top, etc.) tried to resolve a name beyond that
> passwd entry, (or a name that didn't exist), it hung. The process
> stayed around forever and it just kep eating up system time.
Unfortunately, neither of these seems to be the problem in my case.
Thanks to all who responded. Earl (earl@seneca.cs.jmu.edu)
----------------------------------------------------------------------------
richardf@ag.auburn.edu (Richard Feuerriegel)
danny@ews7.dseg.ti.com
ups!kevin@fourx.Aus.Sun.COM (Kevin Sheehan {Consulting Poster Child})
eckhard@ts.go.dlr.de (Eckhard R"uggeberg)
meisner@dlrtcs.da.op.dlr.de (Robert Meisner FE)
carpino@astmib.astro.it (Mario Carpino)
STODDARD@DRAPER.COM (Ike Stoddard)
reagan!wyllys@uu2.psi.com (Wyllys Ingersoll)
miker@il.us.swissbank.com (Mike Raffety)
leisner@eso.mc.xerox.com (Marty Leisner)
odt@dcs.bellcore.com (Dan Transue)
montjoy@thor.ece.uc.EDU (Robert Montjoy)
steve@buick (Steve Swaney)
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:52 CDT