nsupdate to clear MX records REFUSED when current mx rec is invalid

David Jonas djonas at vitalwerks.com
Mon Jul 30 23:42:02 UTC 2007


Mark Andrews wrote:
> There is a test with the wrong sence add vs delete.  It's supposed
> to stop MX records being added without address records and it is
> instead stopping them being deleted.
> 
> Set "check-integrity no;" in named.conf to work around this.
>

That took care of it, Mark. Thanks. I will look at adding the patch
shortly.

>> // Update add seems to work fine.
>>
>> ] nsupdate delete example.com MX
>> ] send
>> update failed: REFUSED
>>
>> // When trying to delete the unresolvable MX record, bind refuses.
>>
>> ] quit
>> mem.c:866: INSIST((((ctx->pools).head == ((void *)0)) ? isc_boolean_true
>> : isc_boolean_false)) failed.
>>
>> // Probably unrelated, but I get that INSIST failure on exit.

Any idea on that mem.c INSIST? It seems it only happens if I give
nsupdate a blank line at some point, with no associated action. It
doesn't seem to cause any problems here, so no big deal. But it seems
like it should probably get closed up. Is there a place I can file it as
(low priority) bug? Er... just realized it could be due to the -O3
CFLAGS. I'll recompile without that when I do the patch.

Thanks again.

Examples:

# nsupdate -y ...
] quit
// no error

# nsupdate -y ...
]
] quit
mem.c:866: INSIST((((ctx->pools).head == ((void *)0)) ? isc_boolean_true
: isc_boolean_false)) failed.
Aborted

# nsupdate -y ...
] server 127.0.0.1
] update add test2.example.com. 120 A 10.1.1.1
]
] quit
// no error

# nsupdate -y ...
]
] server 127.0.0.1
] update add test2.example.com. 120 A 10.1.1.1
]
] quit
mem.c:866: INSIST((((ctx->pools).head == ((void *)0)) ? isc_boolean_true
: isc_boolean_false)) failed.
Aborted


>>
>> ==
>>
>> Bind 9.4.1-P1, RHEL4, From source:
>>  ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var
>>    --mandir=/usr/share/man -with-openssl=/usr/include/openssl
>>    --with-libtool --disable-ipv6 --enable-threads CFLAGS=-O3
>>    -funroll-loops -fomit-frame-pointer
>>
>>



More information about the bind-users mailing list