named died

[Danny Mayer]

		Which one died?  The original machine or the clone?  You may have cloned
  it too perfectly. Or you may have totally unrelated issues.  Is other
software
  running on the machine without problems?

			Danny

At 11:49 PM 10/16/00 +0800, Jonathan Tse wrote:
>
>Thanks Mark,
>
>The thing is very wried to me:
>
>1) Two machines are using the same configurations (both hardware and
>software), indeeded one is "clone" from another one.
>2) only one machines' named behave such a restarting problem. Another dns
>server can up 30 days without problem.
>3) Other than the abnormal named, the whole machine can up for 30 days.
>
>BTW, both machines are using Pentium 200MHz, BSDI4.1 and BIND8.2.2p5
>
>Any hints for troubleshooting?
>
>Best regards,
>Jonathan.
>
>-----Original Message-----
>From: Mark.Andrews at nominum.com <Mark.Andrews at nominum.com>
>To: Jonathan Tse <jonathantse at pacific.net.sg>
>Cc: comp-protocols-dns-bind at moderators.isc.org
><comp-protocols-dns-bind at moderators.isc.org>
>Date: Sunday, October 15, 2000 8:24 PM
>Subject: Re: named died
>
>
>>
>>> Hi,
>>>
>>> I found that my named died after showing this syslog message:
>>>
>>> 14-Oct-2000 03:27:32.1000 insist: critical: ns_main.c:537: INSIST(errno
>==
>>> EINTR): Invalid argument failed
>>>
>>> Can somebody tell me what does it mean?
>>
>> The most common cause is that gettimeofday() returned an
>> invalid time, tv_usec out of range.  This resulted in a
>> out of range value being passed to select/pselect which
>> returned a error.
>>
>> The second possible cause is when the select/pselect complains
>> about a tcp socket that has been closed/reset in the kernel
>> but named has not yet closed.
>>
>> Both of these are the result of kernel bugs.
>>
>> Mark
>>
>>>
>>> Thanks.
>>>
>>> Jonathan.
>>--
>>Mark Andrews, Nominum Inc.
>>1 Seymour St., Dundas Valley, NSW 2117, Australia
>>PHONE: +61 2 9871 4742                 INTERNET: Mark.Andrews at nominum.com
>>
>