[Raw Msg Headers][Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: scheduler died



On Wed, Jun 21, 2000 at 12:30:53PM +0400, Alexey Gadzhiev wrote:

>   Hello,
>    Scheduler silently died after my mailer (invoked via sm) exited with error.
>    Here are last records in scheduler log before stopping:
>   
> 20000620203043 sm: 817651/176//error2 mailer: no such user[exit status 67/67 (addressee unknown)]
> action='failed5.'
> 20000620203454 sm: 817855/173//error2 mailer: mailbox size exceeded.[exit status 69/69 (service unavailable)]
> action='failed5.'
> scheduler: unlink(../queue/Z/h
>                                m@4986): No such file or directory   
> 
> Seems we have  corrupted memory after reading from transport agent in
> scheduler...

I've recently found another another scenario in which the scheduler may die.
Syslogd was for some weirdo reason writing to /dev/console which because
it is a serial interface was blocking over a long time, therefore
all processes which were trying to log something to syslogd via /dev/log
were also blocking after that socket had filled up.  The machine stayed
in this state for several hours until I was able to get to the serial
console which made the system recover.  This scenario happend several times;
each time when the machine was in this state for a longer time which
probably means a some minutes or so the scheduler died.

This was zmailer 2.99.51 running on Redhat 6.1.  Sorry, no core files
available and syslogd's records are obviously not helpful in this
scenario.

  Ralf