On Sat, Oct 17, 2009 at 7:16 PM, Clark Christensen <cdcmi...@yahoo.com> wrote: > Sorry for top-posting... > > What's running on the the server? A Perl CGI script? Apache HTTPD? > mod-perl?
Although I have Apache mod_perl installed, I am running a plain vanilla Perl cgi script for now. > > Is the AJAX exchange asyncronous? Ajax is always asynchronous. That is what the first "A" in Ajax is. > Are you sure the first AJAX exchange is finished when the second one fires? Am I sure? Not really. I guess it is not finished when the second one fires, which is what causes the database lock, no? > Does the AJAX request wait for a 200 response? I am using jQuery. It does what it does. I don't do anything special. > > Assuming Perl, are you explicitly closing the DB with $dbh->disconnect (as > opposed to $dbh = undef)? Does the script end with an exit instruction? I am not doing $dbh->disconnect. DBI is supposed to that automatically when the script finishes. Don't have any issues other than during these Ajax calls. > I'm guessing you're sure there's no writer or writers that jumped-in. I am sure no other process is interfering other than what I have specified. I am sure because I am the only one using my laptop on which I am doing the development. :-) > Are you doing anything like modifying the default 30 second busy_timeout? >Perhaps to too short a timeout? I don't have any timeout set at all. Roger Binns suggestion a busy_timeout, but I am not sure how to even use that. I am using DBI. Where do I set sqlite busy_timeout? Besides, how would that help me? > > Sometimes it's useful to $dbh->disconnect before you print the response. > > I've seen this happen from time to time in my own environment. It's annoying > as hell. It seems to always come down to when and how I disconnect from the > DB file. > > I seldom see it in regular CGI scripts. More often, I see it with > mod_perl::PerlRun (I don't use mod_perl::Registry) when I open the $dbh in a > module where $dbh is exported into the main package. No, as I described, what I think is happening is that two Ajax events are hitting the db at the nearly the same time. The first one is a select (checking the db if the user exists), and the second is an insert (creating a user because the user didn't exist). Even though Javascript has gotten the response from the first event, and started on the second process. sqlite is still busy from the first process, so throws a lockout. Yes, it is very annoying. I have to figure out some other shenanigans to prevent this from happening. > > -Clark > > > > ----- Original Message ---- > From: P Kishor <punk.k...@gmail.com> > To: General Discussion of SQLite Database <sqlite-users@sqlite.org> > Sent: Fri, October 16, 2009 12:53:28 PM > Subject: [sqlite] suggestions for avoiding "database locked" on ajax > > Yes, I know, if it "hurts when I press here," then I shouldn't "press > here," but, so it goes... > > I have an account creation form -- users enter their email and their > desired username, and the form -- > > onblur from the username field, sends off an ajax request to see if > the desired username already exists in the db, and if yes, it sends > back a suggestion; > > onblur from the email field, sends off an ajax request to see if the > email already exists in the db, and if yes... > > If neither the username nor the email exist in the db, then the > application creates a record and informs the user of success. Except, > the previous ajax request (I am assuming it is the previous ajax > request from onblur event from the email field) has locked the > database, and the app returns an error that "the database is locked." > > So, what suggestion might you all have for getting around this? > > > > -- > Puneet Kishor http://www.punkish.org > Carbon Model http://carbonmodel.org > Charter Member, Open Source Geospatial Foundation http://www.osgeo.org > Science Commons Fellow, http://sciencecommons.org/about/whoweare/kishor > Nelson Institute, UW-Madison http://www.nelson.wisc.edu > ----------------------------------------------------------------------- > Assertions are politics; backing up assertions with evidence is science > ======================================================================= > Sent from Madison, WI, United States > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
