INSIST failure and RT #2614

Edward C Clarkson edcclark at
Tue Apr 23 21:56:58 UTC 2002


We ran across the following error yesterday on one of our test servers. 
This occurred shortly after server start; we started the server at 11:37 
or so, and the error occurred about a minute later, after all the zones 
had loaded/notifies sent:

Apr 22 11:38:15.828 general: critical: adb.c:1308: 
INSIST(!adb->cevent_sent) failed

Nothing really special was going on at the time, as far as I can tell from 
the logs at debug 99.  I won't force all that garbage on everyone, but if 
someone thinks that would help, I'd be happy to provide it.  The line 
number in the message probably differs from the regular BIND source b/c of 
our port to z/OS; the failure occurs in the check_exit() 'sanity check' in 

  if (zeroerefcnt) {
        INSIST(!adb->cevent_sent);      /* Sanity check. */
        event = &adb->cevent;
        isc_task_send(adb->task, &event);
        adb->cevent_sent = ISC_TRUE;

I then noticed the following in the 9.2.1rc2 CHANGES file:

1239.   [bug]           Under certain circumstances named could continue 
                        use a name after it had been freed triggering
                        INSIST() failures.  [RT #2614]

This sounds like it -could- be related to the failure we encountered.  I 
was hoping that there was someone that could enough to elaborate on the 
details of RT #2614, and perhaps be so kind as to provide a diff (from 
9.2.0) of the fix?  Many thanks,


Edward C. Clarkson
Email:  edcclark at

More information about the bind-workers mailing list