general/195: SMP enabled kernel causes httpd fatal crash

[Dean Murphy <de...@ant.co.uk>]

	The contract type is `' with a response time of 3 business hours.
	A first analysis should be sent before: Wed Feb 26 10:00:01 PST 1997


>Number:         195
>Category:       general
>Synopsis:       SMP enabled kernel causes httpd fatal crash
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Tue Feb 25 15:40:01 1997
>Originator:     dean@ant.co.uk
>Organization:
apache
>Release:        1.2b7
>Environment:
Dual processor P133 Intel box with 64MBs RAM 
RedHat Linux 4 (build 2.0.18)
gcc 2.0
Postgres 6.0
PHP/FI 2.0.b10
>Description:
If I compile my kernel with SMP support (on a Dual P133 machine) httpd will cause
a complete unrecoverable crash after between 1 and 10 accesses, this is completely
repeatable and stops if I boot a single processor kernel on the same machine.

This is not a temporary hang, the machine does not come back to life and needs a
power cycle or hard reset.

I've tried switching off every optimisation in the kernel in case it's a hardware
clash with no success.

I had a look through your bug database to see if anybody else had similar problems
and noticed that one person, Andrew Korty had a similar problem with an SMP system
and you suggested that it could be dodgy RAM. I've got 64MBs in my machine but
run Linux, NT and 95 without any other problems so I'm not too suspicious of the
RAM.

If I run httpd with the -X option, it will just stop with no output to stdout or
the error_log file.

I've also tried configuring httpd to run only one process, the problem persists.

Help...
>How-To-Repeat:
No idea.
>Fix:
Apart from running with single processor support, no.
%0
>Audit-Trail:
>Unformatted: