[techtalk] Linux/Apache/Webmin question

[Erin Clarke]

Check the path info in the apachectl script (path to httpd
and path to pid file). Could you possibly have two different
versions of apache installed on this machine? Are you running
your site off one httpd.conf or are you using several (as in 
one conf file per site)? I'm assuming you're running from only
one httpd.conf...

You can always restart apache manually by sending a SIGHUP
to the main httpd process (which you can either find by
ps or by cat /var/run/httpd.pid <- or whatever the path to 
the pid file is):

# kill -1 27905 (<- whatever the httpd PID is)
or 
# kill -HUP `cat /var/run/httpd.pid`

Check the server error logs for clues, too. If you're not 
sure where those are, the TransferLog directive in httpd.conf
should contain the correct path. Are the log entries similarly
'backdated'? Any related-looking error messages?

I've never used any apache server admin web interface, so I
am probably not much help there. I've also never run into a 
time problem like the one you are describing, but I have run 
into apachectl/multiple-versions-of-apache problems that had 
some symptoms similar to what you're describing...

I'd be looking at the server admin web interface code, too, 
and any logs it may generate... dunno if I'm much help but 
sometimes more questions lead to ideas on where else to look
for answers.

Erin  8)

On Mon, Oct 02, 2000 at 12:36:34PM -0500, Yvonne Beever wrote:
> I am pretty much a newbie at this, so please excuse my ignorance. We
> have been having some problems with virual hosts using Apache. What we
> discovered this morning is the question I have to ask.  My boss tried to
> stop and restart Apache through the Webmin interface. It stopped, but
> failed on the restart. He went into the command line and issued the
> apachectl start command, which did start it up.  He then looked at the
> process stack and noticed that the httpd processes were dated Sep30, and
> not Oct 2. We were focusing on this dilemma (it appeared that httpd
> really didn't restart) when he discovered that the commands he issued
> right then before checking the process stack again had the date of the
> httpd processes-- Sep 30. We checked the date on the server, and it is
> correct. We then went to our production server, and noticed the same
> thing-- the httpd processes are dated Aug27, and current commands just
> issued also came up as Aug 27. Is there an explanation for this? Is this
> something to be concerned with? He is at a loss to explain this, and of
> course, I know nothing. :-)
> 
> Thanks in advance!
> Yvonne Beever
> Web Services Administrator
> Texas State Library and Archives Commission