check this out...

check this out...

Post by Max Loc » Thu, 04 Feb 1999 04:00:00


Maxwell Lock, Systems Specialist, CRAY Research / SGI UK. 01189 307806

"There are two major products that come out of Berkeley, LSD and UNIX.
 We don't believe this to be a coincidence."       -Jeremy S. Anderson


1. Apache 1.3.9 time outs don't work on SPARC Solaris

Once SIGALRM is delivered and caught in a SPARC Solaris Apache child
process, SIGALRM is blocked forever.

The easiest way to demonstrate this is to only use a very small number of
children and send each one an HTTP 1.1 request. The 1.1 request causes a
keep alive time out to be set. If it's still set to the default, the time
out will trigger at 15 seconds. (You can see the time out trigger if you are
telnetting to port 80.). Once a child receives its second HTTP 1.1 request,
the timeout will not trigger, ever, and the child is hung for all intents
and purposes.

I have searched for this bug in the Apache database and found no mention of
it. Any one else observed this behavior, or can any one else confirm/deny
the behavior? Thanks.

We are seeing this in Apache 1.3.9. The 1.3.12 code does not appear to
contain a fix.

umark -a output:

SunOS xxx 5.7 Generic_106541-10 sun4u sparc SUNW,Ultra-250


2. Serial problem "TIOCMGET failed: Input/output error"

3. bttv: Strange frame drop-outs during TV grabbing

4. Rage With Linux

5. Modem cable pin-outs

6. problem with active ftp and iptables

7. cups: banner appears on print-outs from server only

8. Kernel panic: Attempted to kill the idle task!

9. Keeping Track of my Login/outs?csh script

10. Modem lock outs

11. Can apache/stronghold log when a user 'click outs' of your site?

12. Phony Rev Outs Himself was: Aliyah...

13. Removing banner pages from print outs