On Jan 11, 2007, at 3:18 PM, Mark Stosberg wrote:
Ofer Nave wrote:
Couldn't an option be added as an alternative to db_stack that
indicates
a desire for random ordering, thereby eliminating the need to
precreate
all possible permutations? I'm imaging something like this (modified
version of your example):
$DATABASE::conf{'test'} = {
max_retries => 2,
db_pool => [
[ 'dbi:Sybase:server=prod1;database=test', 'user1', 'pass1',
$attrib ],
[ 'dbi:Sybase:server=prod2;database=test', 'user2', 'pass2',
$attrib ],
[ 'dbi:Sybase:server=prod3;database=test', 'user3', 'pass3',
$attrib ],
], ...
So, db_stack maintains ordering, while db_pool tells DBIx::HA to
randomize the list of servers.
I like the idea of having this feature built-in, but I think the name
could clearer, perhaps call it "load balance":
$DATABASE::conf{'test'} = {
load_balance => 1,
db_stack => [
[ 'dbi:Sybase:server=prod1;database=test','user1','pass1',
$attrib ],
[ 'dbi:Sybase:server=prod2;database=test','user2','pass2',
$attrib ],
], ...
The problem is that there are too many ways to load balance, which
gets even more confusing when you use presistent connections such as
those in mod_perl. Unless you have a large number of client processes
compared the the servers, when you use persistent connections you're
rarely going to be happily load balancing in a way that is equitable.
Certain algorithms are better for certain situations such as a lower
ratio of clients to servers, for example.
And since the developer knows her systems better than I do, let her
write a proper balancing routine for her needs. The trivial ones are
unnecessary for me to include since they're trivial, and the more
advanced ones are too specific to a situation for the general public
(i.e. probably the 3 people using this module) to care about...
As an interesting exercise, I'd be curious to see if you and Ofer can
agree on a load balancing routine to use. :-)
H
