httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mike Rumph <mike.ru...@oracle.com>
Subject Re: [users@httpd] Apache response time slow
Date Fri, 03 Jun 2016 01:09:05 GMT

On 6/2/2016 11:25 AM, Oscar Ithier wrote:
> The typical response was very fast (on miliseconds).
>
> The access to the database is is part of the connection string on the 
> program.
>
> Apache authenticate the user access to the directories that containts 
> the perl programs, using the httpd.conf and the htgroup and 
> htpasswd file.
> ex.
> <Directory /usr/local/apache2/cgi-bin/indocadmin>
> AllowOverride None
> AuthName "INDOC"
> AuthType Basic
> AuthUserFile /usr/local/apache2/conf/.htpasswd
> AuthGroupFile /usr/local/apache2/conf/.htgroup
> <LIMIT GET POST PUT>
> require group INDOCSEC
> order deny,allow
> deny from all
> satisfy any
> </limit>
> </directory>
>
> initialy i get the log message:
>     POST HTTP/1.1 401
>
> then is enter user id and password and click enter.
>
> then i get a delay of 40 seconds or more and the log message:
>   POST HTTP/1.1 200
> indicating OK
>
> No one has changed the Sybase database, i have the only unix 
> workstation, that can telnet to the Database Server.  We are using an 
> intranet classified network.  The Apache, Sybase and the programs are 
> all in the same server.
>
Hello Oscar,

I don't have much idea why you are seeing such a big difference in 
response time.
But I do have some further questions that may be useful to you or to 
someone else on the mailing list that may have a clue what is happening.

First of all, logic would seem to indicate that if we are seeing this 
big of change in the result, then something must have changed in the 
scenario.
A suggested task would be to try to figure out what is different.
Could there be any new programs running on your system that might be 
causing contention of some sort?

If you are using an intranet classified network, does this mean that you 
are using network protocols other than TCP/IP?

Since you are using perl scripts, are you also using mod_perl?

We would like to discern which part of the request/response round trip 
is using up the extra time.
Is the time lost in the Apache httpd server, mod_perl, the perl script 
or in the access to the database?
Would it be possible to run your perl scripts directly from a shell and 
compare the response time to the overall response time through the httpd 
server?

Thanks,

Mike

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Mime
View raw message