BIND 2.3.2-REL on SunSparc running Solaris 7

[Jim Reid]

>>>>> "Crevan" == Crevan Murphy <crevan at dol.ie> writes:

    Crevan> I hope somebody can help me on this one. We are running
    Crevan> BIND 2.3.2-REL on SunSparcs running Solaris 7. There are
    Crevan> two servers acting as slaves to a hidden primary. The
    Crevan> slaves allow recursive lookups. There is nothing special
    Crevan> about their config.  Not infrequently the two slave
    Crevan> servers stop answering recursive queries yet still answer
    Crevan> queries for their authoritative zones. We have found that
    Crevan> the only way to get them back up to speed is to remove the
    Crevan> slave zonefiles and stop then start the process. After
    Crevan> this everything is fine. I had read a thread earlier in
    Crevan> this newsgroup suggesting that a similar problem may be
    Crevan> caused by primaries sending NOTIFYs to the slave but our
    Crevan> logs don't indicate this. in fact there is no untoward
    Crevan> activity indicated in the logs at all! I'm perplexed.  Has
    Crevan> anybody had similar problems? Anybody out there know what
    Crevan> the problem is? How to fix it? Comments?

Could be a bug. File a bug report to bind-bugs at isc.org. If you can get
a core dump from the name server, do that. Keep it in case someone
needs to look at it. Don't mail it in with the bug report.

It would also help if you provided relevant information, like the
named.conf file(s), zone file and so on. And if you can correlate the
hangs to other DNS activity, that would be useful to know too. if the
problem is reproducible, turning the debugging up before the next
incident might help to shed light on the problem.

Note that this discussion is best moved to bind-bugs. Let's not clog
up bind-users with screeds of logs and event traces.