HP3000-L Archives

February 1999, Week 4

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Stan Sieler <[log in to unmask]>
Reply To:
Stan Sieler <[log in to unmask]>
Date:
Thu, 25 Feb 1999 11:55:31 -0800
Content-Type:
text/plain
Parts/Attachments:
text/plain (44 lines)
Paul writes:
> We are running Apache 1.2.5 on our HP3000 with MPE/iX 5.5 we ran into a =
> situation where Apache began taking up 50-75% of the CPU time.  We ended =
> up killing it and restarting it.  We let this go for about 10 mins and =
> we are fairly confident that there was not a good reason for this amount =
> of CPU time to be used.

I think we've seen that in both Apache and the OpenMarket WebServer.
Our solution: abort the server and restart it.

It *might* be associated with a bug I've seen in handling non-byte-stream files
by the byte-stream emulator (maybe), and then again it might not be :)

Have you noticed free disk space disappearing during this time, and heavy
disk activity?  If so, then it might be related.

I haven't had the opportunity/time/inclination to try to track it down
further...but the next time it happens, I'll look more closely.

You might try getting a Debug/iX snapshot of the CPU-eater...

assume it's pin 123 (:showproc job=#j999 will show you the pins)

   logon as manager.sys

   :run httpd.pub.apache; debug
        (or whatever the Apache httpd program name is called on your machine)

When you get the Debug prompt, enter:

   pin #123; tr, i, d

If the stack trace is only one or two lines long, redo the pin/tr command.

Then:

   abort

(this abort the local run of httpd ... not the CPU-eating process)

--
Stan Sieler                                          [log in to unmask]
                                     http://www.allegro.com/sieler.html

ATOM RSS1 RSS2