Mailing List Archive: Linux problem

Linux problem - no response

Aug 22, 1995, 9:06 AM

Post #1 of 4 (483 views)

Forwarded message:
> From stecla@troubador.com Mon Aug 21 16:45:44 1995
> Message-Id: <m0skgwm-000mhzC@aurora.troubador.com>
> Date: Mon, 21 Aug 95 17:12 PDT
> X-Sender: stecla@troubador.com
> Mime-Version: 1.0
> Content-Type: text/plain; charset="us-ascii"
> To: apache-bugs@hyperreal.com
> From: stecla@troubador.com (Stephen Clayton)
> Subject: Re: Bugs 8.10
> X-Mailer: <PC Eudora Version 1.4>
>
> I'm running linux 1.1.59
> Presently running apache 6.5
> downloaded 8.10
> compiled
> During the compile I get 3 warnings:
>
> good compiles...
>
> util.c:Infunction 'get_remote_host'
> util.c:768: warning: passing arg2 of 'pstrdup' discards 'const' from pointer
> target type
> util.c: In function 'find_fqdn':
> util.c:838 warning: passing arg2 of 'pstrdup' discards 'const' from pointer
> target type
>
> more good compiles...
> ...
>
> mod_access.c: In function 'create_access_dir_config':
> mod_access.c:96: warning: assignment makes integer from pointer without a cast
>
> good compiles to end:
>
> moved up the new executable
> tweaked the srm file a bit to get it to run. had to change the addtype line
> a bit
> then it ran fine... but.
> I can't see the server from outside.
> I can't ever connect with the server.
> It just seems to do nothing!
>
> In the error_log the only entry is:
> [Mon Aug 21 17:08:30 1995] Server configured -- resuming normal operations
> then nothing till I reboot the machine or restart apache 6.5 upon which It
> starts running fine again. all other configurations are the same. file
> directories are the same.
>
> Steve
>
> >>
> >> www.troubador.com
> >> stecla@troubador.com
> >>
> >> I'm running apache 6.5 with no problems
> >>
> >>
> >> I can compile 8,10 with linux 1.1.59
> >> but when I try to run it with in place of 6.5 it's like the server just
> >> isn't there. I get no errors on the execution of httpd but IT just does
> >> nothing.
> >
> >If you're using the same config files as before, it ought to work.
> >
> >Nothing in the error-log ?
> >
> >rob
> >
> >
>
>
>
>

Re: Linux problem - no response [ In reply to ]

awm at qosina

Aug 22, 1995, 10:14 AM

Post #2 of 4 (483 views)

Permalink

At 12:37 PM 8/22/95 EDT, you wrote:
>One note --- if this is 0.8.10, it implies something *other* than lock
>file problems as the cause of the hangups (NB that Linux does have lock
>exclusion compiled in for the moment, as I had conflicting reports about
>whether it was needed or not when I first wrote the locks-around-accept()
>code, and decided to play it safe). It would be interesting to attach
>to one of the child processes with gdb and see what they think they're
>doing...
>

Give me a step by step, and I'll get you the output. As I said in my last
message
the regular system works fine. You can check this with http://www.qosina.com
but the virtual doesn't this is at http://206.64.187.60. It just sits
there. If you
telnet to the port, it does respond though.

<Aram>
--
Aram W. Mirzadeh, MIS Manager, Qosina Corporation
http://www.qosina.com/~awm/, awm@qosina.com

Re: Linux problem - no response [ In reply to ]

rst at ai

Aug 22, 1995, 12:37 PM

Post #3 of 4 (475 views)

Permalink

One note --- if this is 0.8.10, it implies something *other* than lock
file problems as the cause of the hangups (NB that Linux does have lock
exclusion compiled in for the moment, as I had conflicting reports about
whether it was needed or not when I first wrote the locks-around-accept()
code, and decided to play it safe). It would be interesting to attach
to one of the child processes with gdb and see what they think they're
doing...

rst

Re: Linux problem - no response [ In reply to ]

rst at ai

Aug 23, 1995, 10:34 AM

Post #4 of 4 (477 views)

Permalink

Date: Tue, 22 Aug 1995 13:14:58 -0400
X-Sender: awm@qosina.com
From: "Aram W. Mirzadeh" <awm@qosina.com>

Give me a step by step, and I'll get you the output.

OK... find the pid of a blocked process (I'm not sure what style of ps
Linux boxes generally come with, but looking for the pid of the server
root process, and grepping the output of a full ps (all processes on
the sytem) looking for those with that number as pid or ppid works for
me --- on SunOS, this is "ps alx | grep <root-pid>").

Once you've identified a process (root or child) which you'd like to
debug, do

% gdb httpd

From within gdb, say

(gdb) attach <pid>

where <pid> is the pid of the process in question. Assuming that
attaching to processes works under Linux, you should then be able to
get a backtrace, examine variables, set breakpoints, etc., with the
usual gdb commands (gdb has a fairly extensive help system; at any
rate, "where" gets a backtrace).

When you want to cut it loose, say "detach", or just quit out of gdb,
which will query for confirmation and then detach it automatically.

rst