Original Question:
I know this is not the proper list to post this
question, but I was wondering if anyone out there has
experienced the same situation as I have:
We have several Netscape web/news servers running,
and often times I find that the httpd will either
stop or get hung, and I've looked at the logs and
don't see anything unusual. What we've resorted to
is to use scripts to monitor the connection to the
web servers and also check to make sure the process
is running and test it throughout the day and recycle
it if it's not responding.
Can anyone suggest to me why this might be happening?
E.g. can intensive I/O on a box cause it to die? I
know that stale NFS mounts will cause it to stop
working. I would welcome any suggestions that you
have.
Summary:
My thanks to the following people who responded:
Rodney Wines <rodney.wines@ahqps.alcatel.fr>
James K Brigman <brigjk@aur.alcatel.com>
Tom Bosick <tbosick@mail.hollywood.com>
Edmundo Farinas <efarinas@mail.sa.omnes.net>
David Evans <DJEVANS@au.oracle.com>
Dave Wanamaker <wanamaker@Radix.Net>
Chris Tubutis <tubutis.chris@tci.com>
>From what I gather, there are many people that have
the same/similar problem and have taken similar
measures to keep the web servers running. I've also
added a summary of what some of them have written:
Edmundo Farinas <efarinas@mail.sa.omnes.net> writes:
It seems to be a bug with Netscape Enterprise/2.0a ).
Moreover, the administration server cuts the logs
files (20-30 Mb) because it shuts down the web server.
"I assume you are using tis fwtk 1.3, possibly
with http-gw 1.4. I understand that the fwtk 2.0
has fixed this problem."
David Evans <DJEVANS@au.oracle.com> writes:
Truss the news process, telnet session, etc. You
will find it's hanging on an "door" call, if I'm
right. This is a new type of semaphore (from the
man pages), `man -k door` to see the calls
involved. This can be traced back to the ncsd
(Name Server Cache Daemon). The quick solution
is to stop and restart the nscd process but this
kills current connections. I've not had a chance
to see why the process locks up. You can totally
disable the nscd processes for a small drop in
connection performance I believe but have not tried
it myself yet.
Dave Wanamaker <wanamaker@Radix.Net>:
We run several Enterprise 3.0 servers using Solaris
2.5 as the OS and we've never had a problem with
the httpd hanging with no reason. We do experience
the httpd not dying or starting when its supposed to
(like stopping the web server through the admin
server) but a kill PID solves that. If you are
running version 3.0 using Solaris 2.5 or 2.5.1, you
will have to install a patch from Solaris. You can
get more information at the following URL:
http://help.netscape.com/faqs/entrprz3x.html
If you're running version 2.0 or older, Netscape has
been to known to have many bugs in the earlier
products.
Chris Tubutis <tubutis.chris@tci.com>:
I think the Enterprise 3.0 server software requires
some adjustment of system parameters - look in the
README for more information; not sure about the
News server software (althogh it *does* want a
*bunch* of memory). I've also heard of this kind
of thing happening with the 2.x servers but don't
remember any specifics. Take a gander at
http://help.netscape.com.
===
Ju
julienlim@rocketmail.com
_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:30 CDT