Re: DBI 1.15+ establishes multiple connections under parent mod_perl process
In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] (Tim Bunce) wrote: > > There were connect() changes made between DBI 1.14 and 1.15 but I'd need > people to look into it for me. Should be trivial to debug by enabling > DBI tracing and Apache::DBi debug. Unless your admin refuses to run any of the mod_perl and Apache::DBI stuff compiled in, because he's a php freak and thinks mod_perl is a resource pig. :\ -- Scott R. Godin| e-mail : [EMAIL PROTECTED] Laughing Dragon Services |web : http://www.webdragon.net/
Re: DBI 1.15+ establishes multiple connections under parent mod_perl process
In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] (Eric Kolve) wrote: > I have traced it back to prepare_cached() (at least that is what I > notice). > > Scott, try replacing your calls on startup with prepare() instead of > prepare_cached(). no, I'm using prepare(). an earlier post thread of mine (php vs perl) has a copy of the script I'm running in it. > I was also able to eliminate the problem if I commented out the > following line in DBI.pm > > # $dbh->STORE('CachedKids', $cache = {}) unless $cache; # line 1021 > sub prepare_cached I haven't tried this though > Of course this is not a solution, but it may give someone else with more > knowledge enough to fix the problem. I will keep digging for answers. -- Scott R. Godin| e-mail : [EMAIL PROTECTED] Laughing Dragon Services |web : http://www.webdragon.net/
Re: DBI 1.15+ establishes multiple connections under parent mod_perl process
What trace level should I use? I figure I can run trace and debug for when it works and doesn't work; dump to two separate log files and post to either the group or just send it to you Tim if you have time to look at it. thanks, --eric Tim Bunce wrote: > > But prepare_cached/prepare has nothing to do with multiple connections. > > There were connect() changes made between DBI 1.14 and 1.15 but I'd need > people to look into it for me. Should be trivial to debug by enabling > DBI tracing and Apache::DBi debug. > > Tim. > > On Fri, Nov 02, 2001 at 07:51:33AM -0800, Eric Kolve wrote: > > I have traced it back to prepare_cached() (at least that is what I > > notice). > > > > Scott, try replacing your calls on startup with prepare() instead of > > prepare_cached(). > > > > I was also able to eliminate the problem if I commented out the > > following line in DBI.pm > > > > # $dbh->STORE('CachedKids', $cache = {}) unless $cache; # line 1021 > > sub prepare_cached > > > > > > Of course this is not a solution, but it may give someone else with more > > knowledge enough to fix the problem. I will keep digging for answers. > > > > --eric > > > > > > "Scott R. Godin" wrote: > > > > > > In article <[EMAIL PROTECTED]>, > > > [EMAIL PROTECTED] (Eric Kolve) wrote: > > > > > > > I think I have found a curious bug in DBI. It seems that since DBI 1.15 > > > > - 1.20, when you bring up apache/mod_perl and execute queries against > > > > the database handle in the parent process (startup.pl), multiple > > > > connections result against the database. If I switch to DBI 1.14, no > > > > such problem occurs. I have found this problem occurs with: > > > > > > > > DBI 1.20 + DBD::Oracle 1.12 > > > > DBI 1.15 + DBD::Oracle 1.07 > > > > DBI 1.16 + DBD::Oracle 1.07 > > > > > > > > > > > > I have turned on Apache::DBI::DEBUG and trace(2) in DBI. Could someone > > > > tell me what I should be looking for or can someone else shed any light > > > > on this? I am not sure if this is necessarily a mod_perl issue or if > > > > mod_perl is just eliciting a bug in DBI. > > > > > > > > thanks, > > > > > > > > --eric > > > > > > I've noticed this too, and it has *seriosly* damaged any credibility I > > > might have gained with the admin I'm up against who is a major PHP > > > proponent, and who refused to even think about installing mod_perl to > > > help the script along after he saw this. :/ > > > > > > -- > > > Scott R. Godin| e-mail : [EMAIL PROTECTED] > > > Laughing Dragon Services |web : http://www.webdragon.net/
Re: DBI 1.15+ establishes multiple connections under parent mod_perl process
But prepare_cached/prepare has nothing to do with multiple connections. There were connect() changes made between DBI 1.14 and 1.15 but I'd need people to look into it for me. Should be trivial to debug by enabling DBI tracing and Apache::DBi debug. Tim. On Fri, Nov 02, 2001 at 07:51:33AM -0800, Eric Kolve wrote: > I have traced it back to prepare_cached() (at least that is what I > notice). > > Scott, try replacing your calls on startup with prepare() instead of > prepare_cached(). > > I was also able to eliminate the problem if I commented out the > following line in DBI.pm > > # $dbh->STORE('CachedKids', $cache = {}) unless $cache; # line 1021 > sub prepare_cached > > > Of course this is not a solution, but it may give someone else with more > knowledge enough to fix the problem. I will keep digging for answers. > > --eric > > > "Scott R. Godin" wrote: > > > > In article <[EMAIL PROTECTED]>, > > [EMAIL PROTECTED] (Eric Kolve) wrote: > > > > > I think I have found a curious bug in DBI. It seems that since DBI 1.15 > > > - 1.20, when you bring up apache/mod_perl and execute queries against > > > the database handle in the parent process (startup.pl), multiple > > > connections result against the database. If I switch to DBI 1.14, no > > > such problem occurs. I have found this problem occurs with: > > > > > > DBI 1.20 + DBD::Oracle 1.12 > > > DBI 1.15 + DBD::Oracle 1.07 > > > DBI 1.16 + DBD::Oracle 1.07 > > > > > > > > > I have turned on Apache::DBI::DEBUG and trace(2) in DBI. Could someone > > > tell me what I should be looking for or can someone else shed any light > > > on this? I am not sure if this is necessarily a mod_perl issue or if > > > mod_perl is just eliciting a bug in DBI. > > > > > > thanks, > > > > > > --eric > > > > I've noticed this too, and it has *seriosly* damaged any credibility I > > might have gained with the admin I'm up against who is a major PHP > > proponent, and who refused to even think about installing mod_perl to > > help the script along after he saw this. :/ > > > > -- > > Scott R. Godin| e-mail : [EMAIL PROTECTED] > > Laughing Dragon Services |web : http://www.webdragon.net/
Re: DBI 1.15+ establishes multiple connections under parent mod_perl process
I have traced it back to prepare_cached() (at least that is what I notice). Scott, try replacing your calls on startup with prepare() instead of prepare_cached(). I was also able to eliminate the problem if I commented out the following line in DBI.pm # $dbh->STORE('CachedKids', $cache = {}) unless $cache; # line 1021 sub prepare_cached Of course this is not a solution, but it may give someone else with more knowledge enough to fix the problem. I will keep digging for answers. --eric "Scott R. Godin" wrote: > > In article <[EMAIL PROTECTED]>, > [EMAIL PROTECTED] (Eric Kolve) wrote: > > > I think I have found a curious bug in DBI. It seems that since DBI 1.15 > > - 1.20, when you bring up apache/mod_perl and execute queries against > > the database handle in the parent process (startup.pl), multiple > > connections result against the database. If I switch to DBI 1.14, no > > such problem occurs. I have found this problem occurs with: > > > > DBI 1.20 + DBD::Oracle 1.12 > > DBI 1.15 + DBD::Oracle 1.07 > > DBI 1.16 + DBD::Oracle 1.07 > > > > > > I have turned on Apache::DBI::DEBUG and trace(2) in DBI. Could someone > > tell me what I should be looking for or can someone else shed any light > > on this? I am not sure if this is necessarily a mod_perl issue or if > > mod_perl is just eliciting a bug in DBI. > > > > thanks, > > > > --eric > > I've noticed this too, and it has *seriosly* damaged any credibility I > might have gained with the admin I'm up against who is a major PHP > proponent, and who refused to even think about installing mod_perl to > help the script along after he saw this. :/ > > -- > Scott R. Godin| e-mail : [EMAIL PROTECTED] > Laughing Dragon Services |web : http://www.webdragon.net/
Re: DBI 1.15+ establishes multiple connections under parent mod_perl process
In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] (Eric Kolve) wrote: > I think I have found a curious bug in DBI. It seems that since DBI 1.15 > - 1.20, when you bring up apache/mod_perl and execute queries against > the database handle in the parent process (startup.pl), multiple > connections result against the database. If I switch to DBI 1.14, no > such problem occurs. I have found this problem occurs with: > > DBI 1.20 + DBD::Oracle 1.12 > DBI 1.15 + DBD::Oracle 1.07 > DBI 1.16 + DBD::Oracle 1.07 > > > I have turned on Apache::DBI::DEBUG and trace(2) in DBI. Could someone > tell me what I should be looking for or can someone else shed any light > on this? I am not sure if this is necessarily a mod_perl issue or if > mod_perl is just eliciting a bug in DBI. > > thanks, > > --eric I've noticed this too, and it has *seriosly* damaged any credibility I might have gained with the admin I'm up against who is a major PHP proponent, and who refused to even think about installing mod_perl to help the script along after he saw this. :/ -- Scott R. Godin| e-mail : [EMAIL PROTECTED] Laughing Dragon Services |web : http://www.webdragon.net/