Very Computer

I recently updated to NCSA 1.5.1 (on SunOS 4), and am having a scad of
troubles with it.

Most importantly, everyone connecting to our site gets the "Document
Contains no Data" message from Netscape whenever they try to follow a
link.  Press the "reload" button, and poof!  Up pops the correct page.

Any ideas?  What could make the server erroneously report no data in a
document once, and then correctly send the document seconds later?
This is becoming more than a little bit of an annoyance....

Thanks in advance for any help you can give.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
R Pickett          *      "Change Change Change" -- Adam Duritz      *


++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

This machine is a Sparc 5, and we're not getting that many hits
simultaneously, so it's not like the server is getting clogged
somehow....  I think....

Please to help....

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
R Pickett          *      "Change Change Change" -- Adam Duritz      *


++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

: I recently updated to NCSA 1.5.1 (on SunOS 4), and am having a scad of
: troubles with it.

: Most importantly, everyone connecting to our site gets the "Document
: Contains no Data" message from Netscape whenever they try to follow a
: link.  Press the "reload" button, and poof!  Up pops the correct page.

: Any ideas?  What could make the server erroneously report no data in a
: document once, and then correctly send the document seconds later?
: This is becoming more than a little bit of an annoyance....

: Thanks in advance for any help you can give.

: ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
: R Pickett          *      "Change Change Change" -- Adam Duritz      *


: ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

We are having the same problems on SunOS 4.1.4 with NCSA HTTPD 1.5.1
and 1.5.2.  Restarting the server seems to fix the problem for a while.

--

  Information Management Branch,
  Natural Resources Canada,                        Voice: (613) 992-2832
  580 Booth St., Ottawa, Ontario, Canada  K1A 0E4  FAX:   (613) 996-2953    

I'm having this problem quite a bit now, with NCSA HTTPD 1.5.?a running
on an Indy R4600.  The question mark is because the source code that I
compiled says it's 1.5.1a, but httpd -v says it's version 1.5.0a;  I'm
assuming it's really 1.5.1a.  

Nothing else is running on the server (some mild mail serving) and our
peak hours only do about 7000 hits.  On campus you can simply reload and
you're fine, but our people across the continent seem to be getting
"document contains no data" after reloading more than 3 times.  I've
noticed that we get more of these complaints when there's some heavy
traffic along the way -- I am definitely not saying that they are
therefore related, tho.  I haven't a clue.  There's a million other
things that could be causing it.  Here's some more info for those
interested:

* My server's parsing ALL html docs, because we want includes in
virtually everything.  Our server load is very light, so processing it
shouldn't be a problem, but this might be triggering a bug;
* I took the suggestion of rebooting the server, so my cron reboots my
server every three hours; I haven't seen any noticeable decline in
complaints or problems;
* We've been seeing this for quite a while, but only recently have
people been unable to merely reload the page.  The first day we got
serious complaints I noticed that the main ISP for our campus T1 was
having difficulties with routers going down.  Related?

I would expect that there are more folks out there seeing this, I hope
that we can pool our information to figure it out.

- kov
--

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

- kov

--

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Quote:> * My server's parsing ALL html docs, because we want includes in
> virtually everything.  Our server load is very light, so processing it
> shouldn't be a problem, but this might be triggering a bug;

        I REALLY need a quick-fix solution, even if it means
"jump back to [insert stable version here]".   PLEASE!

Sincerely,
Bob Donahue

- kov

--

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

What is happening is that the server is core dumping, hence the document
contains no data, and it is generally happening in free or malloc,
because of the server writing to a free'd block of memory, but we're
having a *y hard time finding it.

Quote:>> * My server's parsing ALL html docs, because we want includes in
>> virtually everything.  Our server load is very light, so processing it
>> shouldn't be a problem, but this might be triggering a bug;
>    ditto here
>    I REALLY need a quick-fix solution, even if it means
>"jump back to [insert stable version here]".   PLEASE!

Brandon
--
 Brandon Long           "I think, therefore I am Confused." - RAW
 HTTPd/SDG/NCSA/UIUC  "Here's a nickel, kid.  Get yourself a better computer."

 http://www.veryComputer.com/;Consider myself properly disclaimed.    

Brandon
--
 Brandon Long           "I think, therefore I am Confused." - RAW
 HTTPd/SDG/NCSA/UIUC  "Here's a nickel, kid.  Get yourself a better computer."

 http://www.uiuc.edu/ph/www/blong  Consider myself properly disclaimed.