SQL Server Replication – Will Commands Build Up if Distribution Agent Stops?

Examine the output of sp_helpsubscriptionerrors. One issue I have seen is with DDL commands; they are treated differently in replication so if manual schema changes are being made to the Publisher, they can sometimes fail. There may also be specific DML commands failing on the subscriber due to transactional inconsistencies. If this is the case, you can use the distribution agent profile called "Continue On Data Consistency Errors" which will ignore duplicate key and other errors.

Latency could be causing the problem (according to MS, latency and throughput are the most important factors in transactional replication); you can track that with sp_posttracertoken. Both the Log Reader and Distribution agents have parameters which can be tweaked to compensate for that. Try lowering these values, particularly on the Log Reader.

Distribution Agent (from MSDN):

MaxDeliveredTransactions Is the maximum number of push or pull transactions applied to Subscribers in one synchronization. A value of 0 indicates that the maximum is an infinite number of transactions. Other values can be used by Subscribers to shorten the duration of a synchronization being pulled from a Publisher.

CommitBatchThreshold Is the number of replication commands to be issued to the Subscriber before a COMMIT statement is issued. The default is 1000.

Log Reader Agent (from MSDN):

MaxCmdsInTran Specifies the maximum number of statements grouped into a transaction as the Log Reader writes commands to the distribution database. Using this parameter allows the Log Reader Agent and Distribution Agent to divide large transactions (consisting of many commands) at the Publisher into several smaller transactions when applied at the Subscriber. Specifying this parameter can reduce contention at the Distributor and reduce latency between the Publisher and Subscriber. Because the original transaction is applied in smaller units, the Subscriber can access rows of a large logical Publisher transaction prior to the end of the original transaction, breaking strict transactional atomicity. The default is 0, which preserves the transaction boundaries of the Publisher.

Below script will help you find out how many undistributed commands are there. Also, do you see any errors in sql server or windows logs ? Is it just slow or any failures reported ?

SELECT ( CASE 
           WHEN mdh.runstatus = '1' THEN 'Start - ' + CAST( 
                                         mdh.runstatus AS VARCHAR) 
           WHEN mdh.runstatus = '2' THEN 'Succeed - ' + CAST( 
                                         mdh.runstatus AS VARCHAR) 
           WHEN mdh.runstatus = '3' THEN 'InProgress - ' + CAST( 
                                         mdh.runstatus AS VARCHAR) 
           WHEN mdh.runstatus = '4' THEN 'Idle - ' + CAST(mdh.runstatus AS 
                                                          VARCHAR) 
           WHEN mdh.runstatus = '5' THEN 'Retry - ' + CAST( 
                                         mdh.runstatus AS VARCHAR) 
           WHEN mdh.runstatus = '6' THEN 'Fail - ' + CAST(mdh.runstatus AS 
                                                          VARCHAR) 
           ELSE CAST(mdh.runstatus AS VARCHAR) 
         END )                                   [Run Status], 
       mda.subscriber_db                         [Subscriber DB], 
       mda.publication                           [PUB Name], 
       RIGHT(LEFT(mda.name, Len(mda.name) - ( Len(mda.id) + 1 )), 
       Len(LEFT(mda.name, Len(mda.name) - ( Len(mda.id) + 1 ))) - ( 
       10 + Len(mda.publisher_db) + ( CASE 
                                          WHEN mda.publisher_db = 'ALL' THEN 1 
                                          ELSE Len(mda.publication) + 2 
                                        END ) )) [SUBSCRIBER], 
       CONVERT(VARCHAR(25), mdh.[time])          [LastSynchronized], 
       und.undelivcmdsindistdb                   [UndistCom], 
       mdh.comments                              [Comments], 
       'select * from distribution.dbo.msrepl_errors (nolock) where id = ' + 
       CAST(mdh.error_id AS VARCHAR(8))          [Query More Info], 
       mdh.xact_seqno                            [SEQ_NO], 
       ( CASE 
           WHEN mda.subscription_type = '0' THEN 'Push' 
           WHEN mda.subscription_type = '1' THEN 'Pull' 
           WHEN mda.subscription_type = '2' THEN 'Anonymous' 
           ELSE CAST(mda.subscription_type AS VARCHAR) 
         END )                                   [SUB Type], 
       mda.publisher_db + ' - ' + CAST(mda.publisher_database_id AS VARCHAR) 
                                                 [Publisher DB], 
       mda.name 
       [Pub - DB - Publication - SUB - AgentID] 
FROM   distribution.dbo.msdistribution_agents mda 
       LEFT JOIN distribution.dbo.msdistribution_history mdh 
         ON mdh.agent_id = mda.id 
       JOIN (SELECT s.agent_id, 
                    maxagentvalue.[time], 
                    SUM(CASE 
                          WHEN xact_seqno > maxagentvalue.maxseq THEN 1 
                          ELSE 0 
                        END) AS undelivcmdsindistdb 
             FROM   distribution.dbo.msrepl_commands t (nolock) 
                    JOIN distribution.dbo.mssubscriptions AS s (nolock) 
                      ON ( t.article_id = s.article_id 
                           AND t.publisher_database_id = s.publisher_database_id 
                         ) 
                    JOIN (SELECT hist.agent_id, 
                                 MAX(hist.[time]) AS [time], 
                                 h.maxseq 
                          FROM   distribution.dbo.msdistribution_history hist ( 
                                 nolock) 
                                 JOIN (SELECT agent_id, 
                                              Isnull(MAX(xact_seqno), 0x0) AS 
                                              maxseq 
                                       FROM 
                                 distribution.dbo.msdistribution_history ( 
                                 nolock) 
                                       GROUP  BY agent_id) AS h 
                                   ON ( hist.agent_id = h.agent_id 
                                        AND h.maxseq = hist.xact_seqno ) 
                          GROUP  BY hist.agent_id, 
                                    h.maxseq) AS maxagentvalue 
                      ON maxagentvalue.agent_id = s.agent_id 
             GROUP  BY s.agent_id, 
                       maxagentvalue.[time]) und 
         ON mda.id = und.agent_id 
            AND und.[time] = mdh.[time]