Page 2 of 3

Linux redhat installation problem

Posted: Wed Jan 02, 2002 9:36 pm
by harold@dovesystems
ScriptAlias /cgi-bin/ /home/harold/public_html/cgi-bin/
<Directory "/home/harold/public_html/cgi-bin">
Options ExecCGI
AddHandler cgi-script .cgi .pl
AllowOverride None
Order allow,deny
Allow from all
</Directory>
AddHandler cgi-script .cgi .pl


Thanks!

Harold

Linux redhat installation problem

Posted: Thu Jan 03, 2002 10:56 am
by Kai
Still can't reproduce it here. We'll investigate further if/when we have a RedHat 7.1 box.

Linux redhat installation problem

Posted: Thu Jan 03, 2002 2:46 pm
by harold@dovesystems
Can you suggest a patch in the installation script to remove the check of the CGI URL? I've got Webinator running by copying over my RH7.2 installation, but it is not automatically walking (I probably missed some cron entry the install script does).

Thanks!

Harold

Linux redhat installation problem

Posted: Thu Jan 03, 2002 2:58 pm
by Kai
We were able to test on a RedHat 7.1 machine and could not reproduce the problem. However we're adding a patch that should handle the resulting erroneous "not a full path" issues better. You can do the same by editing the install script and adding double quotes around the second arg to getdir calls. For example, on line 1220, change this:

getdir "CGI URL prefix" $CgiUrl url

to this:

getdir "CGI URL prefix" "$CgiUrl" url

Neither the install script nor Webinator uses cron; Webinator has its own scheduling API via the Texis Monitor process which is started by any Texis or Webinator program.

Linux redhat installation problem

Posted: Thu Jan 03, 2002 3:14 pm
by mark
You should have a process (ps auxw) called "monitor" running. Look in INSTALLDIR/texis/monitor.log to check for monitor errors.

The Webinator admin interface will generally complain if monitor is not running or there is a problem setting the schedule. What did you set the rewalk schedule to?

Linux redhat installation problem

Posted: Thu Jan 03, 2002 4:10 pm
by harold@dovesystems
Actually, the monitor.log looks fine. There were some permission complaints back in November when I was messing with it...

200 Nov 15 06:28:54 Texis Monitor version 04.00.1005676548 starting (pid 2587)
100 Nov 15 06:28:54 Warning: monitor running as root
200 Nov 15 06:28:54 Database Monitor on /usr/local/morph3/texis/testdb/ starting (pid 2589)
006 Nov 15 06:47:24 Can't write to license file /usr/local/morph3/license.new: Permission denied

I probably changed directory permissions or something to get rid of the complaint a little later:

200 Nov 18 02:39:27 Contacting license server
200 Nov 18 02:39:29 Successful reply from license server
006 Nov 18 02:39:29 Can't write to license file /usr/local/morph3/license.new: Permission denied
000 Nov 18 02:39:29 Invalid license /usr/local/morph3/license.key
200 Nov 18 02:39:29 Texis Monitor exiting due to error (pid 2227)
200 Nov 18 06:31:04 Texis Monitor version 04.00.1005676548 starting (pid 13827)
200 Nov 18 06:31:04 License Violation (License Server Contact)
200 Nov 18 06:31:04 Contacting license server
200 Nov 18 06:31:05 Successful reply from license server
200 Nov 18 06:31:06 Database Monitor on /usr/local/morph3/texis/testdb/ starting (pid 13835)

I generally have a different database set to walk each morning at 1am. Here's monitor.log for early this morning:

200 Jan 2 23:21:36 Database Monitor on /usr/local/morph3/texis/FccOmd/db1/ exiting (pid 28575)
200 Jan 3 01:57:21 Database Monitor on /usr/local/morph3/texis/FccRules/db1/ starting (pid 28709)
200 Jan 3 02:02:21 Database Monitor on /usr/local/morph3/texis/FccRules/db1/ exiting (pid 28709)
200 Jan 3 02:06:51 Database Monitor on /usr/local/morph3/texis/FccRules/db1/ starting (pid 28750)
200 Jan 3 02:11:51 Database Monitor on /usr/local/morph3/texis/FccRules/db1/ exiting (pid 28750)

Looking at my query log, the times correspond to times people have done searches. I see nothing about an attempt to start a walk.

There should have been a walk of FccCsb this morning at 1am. The last walk was one manually started by me:

Last complete walk: 2002-01-01 19:11:08 (took 53 minutes 10 seconds)
Success. 6,628 pages (166,056,439 bytes) (224 errors) (165 duplicates)

Perhaps the script realizes it has not been a week since the site was walked and therefore decided to not walk it this morning? Maybe I just have to wait a week...

Ideas?

You guys are great! Especially since I'm still using the free version. I'll try to squeeze the money together soon for the commercial version.

Thanks!

Harold

Linux redhat installation problem

Posted: Thu Jan 03, 2002 4:52 pm
by mark
Recentness of walk is irrelevant to the scheduler. It walks when scheduled. What are the actual settings of your "Rewalk Type" and "Rewalk Schedule" "Frequency" and "Hour"?

Check your "Walk Status" for evidence of trying to rewalk.

If there's no evidence of attempts, try changing the schedule to something else, click update (not go), then change it back and update (not go) again. Maybe there was a problem setting the schedule the first time and you didn't notice the warning/error message.

Linux redhat installation problem

Posted: Thu Jan 03, 2002 5:50 pm
by harold@dovesystems
Looking at FccCsb (the one that should've run at 1 this morning) walk status, we find:

Webinator Walk Report for FccCsb

Creating database /usr/local/morph3/texis/FccCsb/db1...Done.
Walk started at 2002-01-01 18:17:08 (by user)

rewalk type is new
rewalk schedlue is every Thursday.

I'll go ahead and try changing and updating each of the walks and see what happens tonight.

Thanks!

Harold

Linux redhat installation problem

Posted: Thu Jan 03, 2002 6:24 pm
by mark
What about "Hour"? That should be set to "1 AM".

Linux redhat installation problem

Posted: Thu Jan 03, 2002 6:34 pm
by harold@dovesystems
Yes, it is set to 1am. Sorry I didn't mention that. I'll reschedule everything tonight and see what happens.

Thanks!

Harold