On 4/11/02 at 9:54 AM, [EMAIL PROTECTED] (David Shrewsbury) wrote: > Hey gang. Couldn't find an answer to this in the archives. We have a > DB2 database that we access via mod_perl scripts. We have been > getting errors in the Apache log files every morning whenever we > first try to access the database. The error is: > > [Thu Apr 11 09:09:49 2002] null: DBD::DB2::db selectall_arrayref failed: > [IBM][C > LI Driver] CLI0108E Communication link failure. SQLSTATE=40003 at > /usr/local/ap > achessl/perl/reports/trans_history.pl line 90 > > It takes a restart of the web server to eliminate this problem. > Is this the "Morning Bug" that I am encountering? I looked at the > DB2.pm module and it appears that the ping() method has been > implemented so I would think that this would prevent the "Morning > Bug" from showing up. Should I reimplement the ping() method > according to the Apache::DBI manpage? > > -David >
I have many internal systems which use Apache::DBI and DBD::DB2. They're often up for months without being restarted, and I don't use ping(). So, it is possible. Are you the DBA too? A possibility is that the DB is being brought down into offline mode during the night for backups. That'd kill your cached connections. I'll send a fuller answer when I have a chance to reexamine our set up. Bill