Module: ActiveFedora::Sharding::ClassMethods

Defined in:
lib/active_fedora/sharding.rb

Instance Method Summary collapse

Instance Method Details

#assign_pid(obj) ⇒ String

if you are doing sharding, override this method to do something other than use a sequence

Returns:

  • (String) —

    the unique pid for a new object



43
44
45
46
47
48
49
50
51
52
53
54
55
 
# File 'lib/active_fedora/sharding.rb', line 43

def assign_pid(obj)
  args = {}
  args[:namespace] = obj.namespace if obj.namespace
  # TODO: This juggling of Fedora credentials & establishing connections should be handled by 
  # an establish_fedora_connection method,possibly wrap it all into a fedora_connection method - MZ 06-05-2012
  if ActiveFedora.config.sharded?
    credentials = ActiveFedora.config.credentials[0]
  else
    credentials = ActiveFedora.config.credentials
  end
  fedora_connection[0] ||= ActiveFedora::RubydoraConnection.new(credentials)
  fedora_connection[0].connection.mint(args)
end

#connection_for_pid(pid) ⇒ Rubydora::Repository

Uses #shard_index to find or create the rubydora connection for this pid

Parameters:

  • pid (String) —

    the identifier of the object to get the connection for

Returns:

  • (Rubydora::Repository) —

    The repository that the identifier exists in.



14
15
16
17
18
19
20
21
22
23
24
 
# File 'lib/active_fedora/sharding.rb', line 14

def connection_for_pid(pid)
  idx = shard_index(pid)
  unless fedora_connection.has_key? idx
    if ActiveFedora.config.sharded?
      fedora_connection[idx] = RubydoraConnection.new(ActiveFedora.config.credentials[idx])
    else
      fedora_connection[idx] = RubydoraConnection.new(ActiveFedora.config.credentials)
    end
  end
  fedora_connection[idx].connection
end

#shard_index(pid) ⇒ Integer

This is where your sharding strategy is implemented – it’s how we figure out which shard an object will be (or was) written to. Given a pid, it decides which shard that pid will be written to (and thus retrieved from). For a given pid, as long as your shard configuration remains the same it will always return the same value. If you’re not using sharding, this will always return 0, meaning use the first/only Fedora Repository in your configuration. Default strategy runs a modulo of the md5 of the pid against the number of shards. If you want to use a different sharding strategy, override this method. Make sure that it will always return the same value for a given pid and shard configuration.

Returns:

  • (Integer) —

    the index of the shard this object is stored in



33
34
35
36
37
38
39
 
# File 'lib/active_fedora/sharding.rb', line 33

def shard_index(pid)
  if ActiveFedora.config.sharded?
    Digest::MD5.hexdigest(pid).hex % ActiveFedora.config.credentials.length
  else
    0
  end
end