Re: Using fork with DBI to create simultaneous db connections.

In general, you can't use a DBH across a fork. I don't know of any db clients where the handle will survive a fork.

You write, however,

it appears that all the children end up sharing the same DBH

How did you check for that? The good way to check would be to have your program fork, let each child open a connection, then sleep for a minute or something. During that time, log on to the database using another tool (Toad or DB-Artisan or something) and see how many connections are in your name. If the numbers agree, you're done.

Do not rebuke them with harsh words ... but rather lead them gently - with URLs - so that they may learn wisdom.

Comment on Re: Using fork with DBI to create simultaneous db connections. Download Code

Replies are listed 'Best First'.
Re^2: Using fork with DBI to create simultaneous db connections. by thor (Priest) on Feb 08, 2005 at 04:26 UTC
The good way to check would be to have your program fork, let each child open a connection, then sleep for a minute or something. During that time, log on to the database using another tool (Toad or DB-Artisan or something) and see how many connections are in your name. If the numbers agree, you're done. ...or, you could have each child just print the database handle. Like all perl objects, it's just a reference, so it should print something like `DBI::db=HASH(0x1223456)`. If they all print the same reference, done. No monkeying around on the database at all (well, except the DBI calls that is...but you already knew that :P). thor `Feel the white light, the light within Be your own disciple, fan the sparks of will For all of us waiting, your kingdom will come`	[reply]
Re^3: Using fork with DBI to create simultaneous db connections. by etcshadow (Priest) on Feb 08, 2005 at 05:34 UTC
Actually, that's incorrect, and it's probably the same mistake that the original poster made. The flaw is that the memory address shown as part of the stringification of the reference is the relative memory address to that process. The operating system uses something called "virtual memory", which abstracts the actual physical memory location away from your processes. In truth, the physical memory addresses may change from time to time, as your process is swapped and paged and so on. Anyway, here's a quick demo: `prompt> perl -le 'my $pid = fork; my $x = [$pid]; print "$x->[0] -- $x +"' 0 -- ARRAY(0x80d11ec) 24694 -- ARRAY(0x80d11ec) prompt>` [download] You can see that each process has it's own array, with the reference stored in $x, yet in both processes, the memory address appears the same. `------------ :Wq Not an editor command: Wq` [download]	[reply] [d/l] [select]
Re^4: Using fork with DBI to create simultaneous db connections. by thor (Priest) on Feb 08, 2005 at 12:46 UTC
Hell...I learned something. Not that that's bad...I just don't like being wrong. :) thor `Feel the white light, the light within Be your own disciple, fan the sparks of will For all of us waiting, your kingdom will come`	[reply]