restart - problem

[Joseph S D Yao]

> we are running several nameservers with bind 8.1.2 on solaris 2.6 for a
> few month.
> Six times a day we are reloading them.
> Everything was alright till we made a restart. Now we have the strange
> problem, that some of them don't work any more, even all have the same
> configuration (the same named.conf). If you ask one of them, a time out
> is all you get.

Try sending that one some signal - to get a cache dump, or re-load the
server files, or anything.

There have been reports of servers hanging trying to query specific
other servers.  A signal [a real Unix signal vs. a fifo message] will
cause such a "hung" read to break.

I think someone posted a timeout patch here recently ... or mentioned
that 8.2 had it ...

--
Joe Yao				jsdy at cospo.osis.gov - Joseph S. D. Yao
COSPO/OSIS Computer Support					EMT-B
-----------------------------------------------------------------------
	PLEASE ... send or Cc: all "COSPO/OSIS Computer Support"
		     mail to sys-adm at cospo.osis.gov
-----------------------------------------------------------------------
      This message is not an official statement of COSPO policies.