The following reply was made to PR general/1680; it has been noted by GNATS.

From: Marc Slemko <[EMAIL PROTECTED]>
To: Apache bugs database <[EMAIL PROTECTED]>
Cc:  Subject: Re: general/1680: Critically anomalous behavior when upgrading   
from 1.3a1 to 1.3b3 (fwd)
Date: Sat, 17 Jan 1998 16:45:55 -0700 (MST)

 ---------- Forwarded message ----------
 Date: Sat, 17 Jan 1998 02:56:22 -0800 (PST)
 From: Morgan Davis <[EMAIL PROTECTED]>
 To: Marc Slemko <[EMAIL PROTECTED]>
 Subject: Re: general/1680: Critically anomalous behavior when upgrading   from 
1.3a1 to 1.3b3
 
 Marc Slemko writes:
 > > Keep in mind that 1.3a1 is using the exact same settings as was 1.3b3, and
 > > it doesn't exhibit this behavior.
 > 
 > Yes, but USE_FLOCK_SERIALIZED_ACCEPT was broken with 1.3a1.  It was fixed
 > in b2.
 > 
 > Comment out the USE_FLOCK_SERIALIZED_ACCEPT bit from the section
 > appropriate to your OS in src/main/conf.h. 
 
 Marc, I've recompiled without the above #define and just restarted the
 server with fingers crossed.  I'm calling this the "modified 1.3b3".
 
 > Are you _sure_ the child is a zombie?  If the parent is dead, it shouldn't
 > be.  What releveant processes and info does a ps -axlw show? 
 
 Yes, it was a zombie.  Could not kill it with a -9, no matter what,
 and it's parent PID was 1.
 
 When I was running the "virgin" 1.3b3, top indicated that most of the
 httpds were in a "lockf" state.  With the modified 1.3b3, most seem to
 be in select, accept, or sbwait states, ala 1.3a1.  Haven't seen a
 lockf yet.
 
 > Nothing in the error log?
 
 Here are just the interesting items (this is the virgin version):
 
 ----------------------------------------
 [Thu Jan 15 09:12:58 1998] [notice] Apache/1.3b3 configured -- resuming normal 
operations
 [Thu Jan 15 09:15:02 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 09:20:28 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 09:20:40 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/king/oasis/index.html
 [Thu Jan 15 09:25:30 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 09:26:59 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/king/support/onair/index.html
 [Thu Jan 15 09:27:01 1998] [error] Premature end of script headers: 
/web/king/prodcomp/cgi-bin/prodcomp/count.cgi
 [Thu Jan 15 09:30:33 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 09:31:35 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/king/gibkeeg/index.html
 [Thu Jan 15 09:35:36 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 09:40:38 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 09:45:41 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 09:50:43 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 09:52:32 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/king/oasis/index.html
 [Thu Jan 15 09:55:26 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/crash/jeffmj/index.html
 [Thu Jan 15 09:55:46 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 10:00:49 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 10:05:52 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 10:10:55 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 10:15:57 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 10:18:27 1998] [error] server reached MaxClients setting, consider 
raising the MaxClients setting
 [Thu Jan 15 10:19:08 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/king/accessc/index.html
 [Thu Jan 15 10:20:59 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 10:21:55 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/king/oasis/index.html
 [Thu Jan 15 10:25:41 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/king/oasis/index.html
 [Thu Jan 15 10:26:11 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 10:27:25 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/king/oasis/index.html
 [Thu Jan 15 10:31:49 1998] [crit] (0)Undefined error: 0: mmap_handler: mmap 
failed: /web/support/cts/index.html
 [Thu Jan 15 10:32:35 1998] [error] child process 2335 still did not exit, 
sending a SIGKILL
 [Thu Jan 15 10:32:52 1998] [error] could not make child process 2335 exit, 
attempting to continue anyway
 [Thu Jan 15 10:32:52 1998] [notice] httpd: caught SIGTERM, shutting down
 bind: Address already in use
 httpd: could not bind to port 80
 bind: Address already in use
 httpd: could not bind to port 80
 bind: Address already in use
 httpd: could not bind to port 80
 ----------------------------------------
 
 This run shows the short-lived server before I put 1.3a1 back up.  The
 last few lines were due to the zombie child not going away.
 
 Now running the "modified" 1.3b3, I still see the "mmap_handler" error
 messages showing up.  I see one appear in the log about every few
 minutes (and right now the server is only taking about 100 hits a
 minute as it's 2:50AM).  What's that mean?
 
 > Are you sure you aren't running low on file descriptors?  Unlikely
 > anything should change between a1 and b3, but...
 
 Pretty sure.  /etc/login.conf has all the resources jacked way up --
 shouldn't be a problem.
 
 I'll let this modified 1.3b3 run for now until it acts up (hopefully
 it won't).  Are there any kinds of lines in the error log you're
 particularly interested in?  There's a lot of junk (File not found,
 etc.) which is why we normally run it with error logging directed to
 /dev/null.
 
 --Morgan
 
    ___ ____ ___
   /     /  /__   Network  |  619/637-3600 x 1020  |  mailto:[EMAIL PROTECTED]
  /__   /  ___/  Services  |  FAX 619/637-3630     |  http://www.cts.com/
 

Reply via email to