Difference between revisions of "Design Specs: Database Connectivity"
(→Plans) |
|||
Line 65: | Line 65: | ||
setting up attributes such as hosts, user, password, caching for AceDB ect through Configuration file. | setting up attributes such as hosts, user, password, caching for AceDB ect through Configuration file. | ||
+ | |||
+ | Catalyst::Plugin::ConfigLoader | ||
==Things to think about== | ==Things to think about== |
Revision as of 11:11, 22 February 2010
Contents
Objectives
Part of the WormBase::API. Needs to be able to connect to multiple databases (multiple instances of the same database in some cases). Should be fault tolerant if a server or database crashes.
Plans
http://bitbucket.org/tharris/wormbase/src/tip/lib/WormBase/API/Role/Service.pm a Moose Role module which defines the common attributes and behavior for different databases.
eg.
'hosts' array reference attribute which has all the available hosts from configuration file.
reconnect: abstract method for the database connection, fail-over is also implementation here:
- use Moose 'around' method modifier to check on the db handle every time it is accessed
around 'dbh' sub {}
Note that attribute accessors also count as methods for the purposes of satisfying the requirements of a role.
- do "reconnect" if dbh is undef
- will try to connect for maximum $ReconnectMaxTries(now this is hard coded as 5, will replace by conf) times
- host is picked randomly among the pool
my $host = $self->hosts->[ rand @{$self->hosts} ];
This Role is consumed by each Database Classes(Service module)
- http://bitbucket.org/tharris/wormbase/src/tip/lib/WormBase/API/Service/acedb.pm
- http://bitbucket.org/tharris/wormbase/src/tip/lib/WormBase/API/Service/gff.pm
each has its own connection method
AceDB use Ace; sub connect { my $self = shift; my ($host,$port,$user,$pass)=@_; return Ace->connect(-host => $host, -port => $self->port, -user=>$self->user, -pass=>$self->pass, ); # @cache); } GFF database use Bio::DB::GFF; sub connect { my $self = shift; my ($host,$port,$user,$pass)=@_; return Bio::DB::GFF->new( -user => $user, -pass => $pass, -dsn => "dbi:mysql:database=".$self->species.";host=" . $host, ); }
Question
When Ace time out would the dbh becomes undef? How useful is the Ace reopen method? For now I just re-connect.
Continued Work
setting up attributes such as hosts, user, password, caching for AceDB ect through Configuration file.
Catalyst::Plugin::ConfigLoader
Things to think about
persistent connections
Apache::DBI DBI connect_cached DBIx::Class
Also, because Apache::DBI's connection caching never checks to see if the process and/or thread id changed, it will give you an invalid $dbh from a previous process/thread if you use a connection, then thread/fork, then try to re-connect to the same source, which is a Bad Thing.
DBIx::Class already manages its connections for you, and therefore it cannot benefit from Apache::DBI under any scenario. It makes one connection per-process, and keeps that connection persistent, reconnecting only if the connection appears to have died, or if you fork/thread over to another process/thread-id. The only Apache::DBI issue in DBIx::Class is that Apache::DBI will actually thwart DBIx::Class's connection management code, and cause it to use the same (and invalid) connection in a new process, in cases such as (as mentioned above) if you make a DBI connection before forking in a prefork mod_perl server.
Resources
The current connectivity code is part of the API. The API, in turn, is inside of the lib directory of the WormBase mercurial repository:
http://bitbucket.org/tharris/wormbase/src/tip/lib/WormBase/API/Service/
Currently, each database is a "Service". When the API is instantiated, connections are made to each required service. Database handles are cached. When a database handle is needed, the service first checks to see if it is live, if not it refreshes the connection. This would probably be the ideal location to implement failover -- check the current connection, if dead, try a different database server.
- Note that attribute accessors also count as methods for the purposes of satisfying the requirements of a role.
http://bitbucket.org/tharris/wormbase/src/9009075c4c5a/lib/WormBase/API/Role/Service/gff.pm
around 'dbh' => sub { my $orig = shift; my $self = shift; my $species = $self->species; # Do we already have a dbh? HOW TO TEST THIS WITH HASH REF? if ($self->has_dbh) { $self->log->debug(" gff-dbh for $species exists and is alive!"); return $self->$orig; } else { $self->log->debug(" gff-dbh for $species doesn't exist yet; trying to connect"); my $dbh = $self->connect($species); } };
Note also that these Services use Moose "Roles" to define shared methods and variables like connect().
Here's the acedb service:
http://bitbucket.org/tharris/wormbase/src/tip/lib/WormBase/API/Service/acedb.pm
The acedb service "consumes" the generic Service role:
http://bitbucket.org/tharris/wormbase/src/9009075c4c5a/lib/WormBase/API/Role/Service.pm
This nomenclature and structure is a bit confusing and probably needs to be streamlined/refined!!