Click here to monitor SSC
SQLServerCentral is supported by Red Gate Software Ltd.
 
Log in  ::  Register  ::  Not logged in
 
 
 
        
Home       Members    Calendar    Who's On


Add to briefcase

distribution database maintenance Expand / Collapse
Author
Message
Posted Thursday, October 31, 2013 8:16 AM
SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: Tuesday, January 21, 2014 8:49 AM
Points: 116, Visits: 713
How often we need perform maintenance tasks on distribution database? As the data is continuously populated and deleted from the distribution tables, the fragmentation on most of the indexes will be hitting above 90 percent very frequently. Any recommendations ?
Post #1510199
Posted Thursday, October 31, 2013 9:06 AM
Ten Centuries

Ten CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen Centuries

Group: General Forum Members
Last Login: Wednesday, April 16, 2014 2:08 AM
Points: 1,350, Visits: 15,200
It depends a little on how much data is building up and if its causing an issue. In a high throughput distribution db you can have the replication cleanup jobs running every few minutes to keep the data levels low (assuming the distribution agent is doing its job).

With low data levels index fragmentation isn't a problem.
Post #1510249
Posted Thursday, October 31, 2013 9:46 AM
SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: Tuesday, January 21, 2014 8:49 AM
Points: 116, Visits: 713
MysteryJimbo (10/31/2013)
It depends a little on how much data is building up and if its causing an issue. In a high throughput distribution db you can have the replication cleanup jobs running every few minutes to keep the data levels low (assuming the distribution agent is doing its job).

With low data levels index fragmentation isn't a problem.


Very min data is retained in the distribution database. Recently we are continuously experiencing blocking(>30 mins) on distribution server which in turn causing latency. Trying to figure out exactly what's causing this issue?

the most common causes of blocking are with distribution clean up job blocking the update statements on msdistributin_history table and the rest related to msrepl_commands table.

Post #1510278
Posted Thursday, October 31, 2013 10:25 AM
Ten Centuries

Ten CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen Centuries

Group: General Forum Members
Last Login: Wednesday, April 16, 2014 2:08 AM
Points: 1,350, Visits: 15,200
That is pretty common and is down to some barely acceptable code (in my opinion) in the clean up procs. In many cases some of this code isn't even required. I've seen the same job issues many times over, in many environments.

I resolved the deadlock issues in my environment by creating my own versions of the MS procs. It's generally not a problem with distribution database itself.
Post #1510316
Posted Thursday, October 31, 2013 10:50 AM
SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: Tuesday, January 21, 2014 8:49 AM
Points: 116, Visits: 713
I appreciate if you can provide more information on the custom stored procs.
Post #1510333
« Prev Topic | Next Topic »

Add to briefcase

Permissions Expand / Collapse