Author: Chris Skorlinski [MSFT]

ReplTip – Publishing 1 Article into 2 Publications bloats Distribution DB

Chris Skorlinski – Microsoft SQL Server Escalation Services While visiting a customer site, we discussed consequences of publishing common article\tables into multiple Publications. For this customer, each Publication contains same set of core or common tables used by all subscribers, but then some subscribers had tables unique to just that subscriber. For example, all Publications
Read More »

ReplTip – Distribution cleanup Job running but shows zero rows deleted

  By Taiyeb Zakir, Microsoft SQL Escalation Services Consider the following scenario: Distribution cleanup Job is running and you see that number of rows in msrepl_commands and msrepl_transactions are reducing but the Cleanup Job history says 0 rows deleted: For example: Removed 0 replicated transactions consisting of 0 statements in 12961 seconds (0 rows/sec). [SQLSTATE
Read More »

Troubleshooting ‘The process could not execute ‘sp_repldone/sp_replcounters’

Chris Skorlinski, SQL Server Escalation Services I worked on customer’s Transactional Replication issue today where LogReader was failing, restarting, then failing again.  We used link below to enable logging to get a clearer picture of failure.  http://blogs.msdn.com/b/repltalk/archive/2010/03/04/kb-article-312292-how-to-enable-replication-agents-for-logging-to-output-files-in-sql-server.aspx 16:02:45 OLE DB DISTOLE: sp_MSget_last_transaction 16:02:45 Publisher: {call sp_repldone ( 0x000a35e7000315e7005b, 0x000a35e7000315e7005b, 0, 0)} 16:32:45 Status: 2, code:
Read More »

@alt_snapshot_folder for PULL Subscribers

To improve SQL Distribution agent performance when applying Snapshot, first ZIP then copy the snapshot files locally to the subscriber. This is a great solution for subscribers on WAN international networks as ZIP can be smaller and faster to copy then Replication can move uncompressed Snapshot files.  Testing shows a 1gb file can take >
Read More »