Alan Fiebig wrote:

> When smtprcv is running, if you check taskmanager and have the 'Threads'
column
> enabled to display, you will see many copies (threads) of smtprcv running,
one
> for each concurrent connection to the service.
>
> When smtprcv is stopped, either from the service manager, from the netsvc
command,
> the kill command, or any other means, you MUST wait until every thread has
stopped
> and smtprcv is no longer listed in taskmanager before restarting the
service.
>
> Failure to do so will result in exactly what you are describing above... a
service
> that is neither running nor stopped, nor that can be recovered from. The
server
> typically needs to be rebooted to correct it.

Humm. In my case, the condition happened when the system was unattended. My
restart script runs at 5 am, and it was dead at 12:30 am, so I doubt that
was the cause of the lockup.

> I have seen it take as long as 45 seconds for all the threads to expire,
and when
> my monitoring system detected smtprcv had failed and tried immediatly
restarting it,
> I often ended up with this condition. I now have a 60 second delay built
in between
> stopping the service and trying to restart it.

I'll look at the script anyway. But hopefully this is something that we
won't need in the future...starting and stopping these things should only be
done in emergencies, not as a matter of course.

            Randy.


This is the discussion list for the IMS Free email server software.
  To unsubscribe send mailto:[EMAIL PROTECTED]

            Delivered by Rockliffe MailSite
           http://www.rockliffe.com/mailsite
                Rock Solid Software (tm)

Reply via email to