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 ««123»»

Transactional Replication Performance Issues After Migration From 2000 to 2008 Expand / Collapse
Author
Message
Posted Thursday, November 14, 2013 6:25 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Friday, November 22, 2013 2:40 PM
Points: 16, Visits: 71
Distribution Agent Log
************************ STATISTICS SINCE AGENT STARTED ***********************
11-14-2013 13:20:45

Total Run Time (ms) : 394605 Total Work Time : 389457
Total Num Trans : 5194 Num Trans/Sec : 13.34
Total Num Cmds : 8928 Num Cmds/Sec : 22.92
Total Idle Time : 0

Writer Thread Stats
Total Number of Retries : 0
Time Spent on Exec : 25784
Time Spent on Commits (ms): 1622 Commits/Sec : 0.13
Time to Apply Cmds (ms) : 389457 Cmds/Sec : 22.92
Time Cmd Queue Empty (ms) : 157 Empty Q Waits > 10ms: 10
Total Time Request Blk(ms): 157
P2P Work Time (ms) : 0 P2P Cmds Skipped : 0

Reader Thread Stats
Calls to Retrieve Cmds : 2
Time to Retrieve Cmds (ms): 369629 Cmds/Sec : 24.15
Time Cmd Queue Full (ms) : 19843 Full Q Waits > 10ms : 128
Post #1514246
Posted Thursday, November 14, 2013 6:40 AM
Ten Centuries

Ten CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen Centuries

Group: General Forum Members
Last Login: Monday, August 18, 2014 3:04 AM
Points: 1,362, Visits: 15,269
It looks like the delivery of commands is whats taking the time. Have you compared the distribution agent profile settings between the servers?
Post #1514250
Posted Thursday, November 14, 2013 7:49 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Friday, November 22, 2013 2:40 PM
Points: 16, Visits: 71
Only differences are BCPBatchSize and QueryTimeout which we haven't changed and PollingInterval which changed to 5 by default and we have changed back to 10
Post #1514293
Posted Thursday, November 14, 2013 8:04 AM
Ten Centuries

Ten CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen Centuries

Group: General Forum Members
Last Login: Monday, August 18, 2014 3:04 AM
Points: 1,362, Visits: 15,269
chris.roddis-ferrari (11/14/2013)
Only differences are BCPBatchSize and QueryTimeout which we haven't changed and PollingInterval which changed to 5 by default and we have changed back to 10


None of those would make any difference to command delivery. Have you checked for blocking on distribution db and subscription db?

Checked latency using a tracer token?


These are the parameters which modify delivery rate.

[-CommitBatchSize commit_batch_size]
[-CommitBatchThreshold commit_batch_threshold]
[-MaxDeliveredTransactions number_of_transactions]
[-PacketSize packet_size]
[-SubscriptionStreams [1|2|...64]]
Post #1514300
Posted Thursday, November 14, 2013 8:36 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Friday, November 22, 2013 2:40 PM
Points: 16, Visits: 71
Tracer Token show Publisher to Distributor (same box) as a couple of secs and all the latency is Distributor to Subscriber - 12 minutes in the one I just did. There is no blocking on the distribution db, just a ASYNC_NETWORK_IO wait of about 2 secs on the distribution process. There has always been a level of blocking on the subscriber even before the upgrades started, but this has never caused performance issues. I am not able/don't know how to tell if this has increased
Post #1514320
Posted Thursday, November 14, 2013 9:03 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Friday, November 22, 2013 2:40 PM
Points: 16, Visits: 71
We have also just expanded tempDB on the subscriber server to have 8 data files to match the number of processors.
Post #1514335
Posted Thursday, November 14, 2013 9:22 AM
Ten Centuries

Ten CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen Centuries

Group: General Forum Members
Last Login: Monday, August 18, 2014 3:04 AM
Points: 1,362, Visits: 15,269
I'm assuming network speed/bandwidth has been eliminated since it was a SQL upgrade.

So this may be down to differences in functionality between the 2000 distribution agent and 2008. Maybe execution of the repl procs on the subscriber is slow due to blocking or poor plan.

Did you run sp_updatestats post subscriber upgrade?
Are you able to trace the sp_MSins/upd/del procs to see their execution and test speed?

You could also look in the procedure cache, its possible the 2000 agent has a different plan to 2008.
Post #1514351
Posted Friday, November 15, 2013 1:40 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Friday, November 22, 2013 2:40 PM
Points: 16, Visits: 71
Network hasn't changed so that was ruled out.
Stats were updated for the MSreplication_subscriptions table.
We will do some comparing of the repl procs today to see if that shows anything.

We also have the DBs in 2000 compatibility mode still and have changed a few to 2008 but that doesn't appear to have resolved
Post #1514624
Posted Friday, November 15, 2013 2:08 AM
Ten Centuries

Ten CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen CenturiesTen Centuries

Group: General Forum Members
Last Login: Monday, August 18, 2014 3:04 AM
Points: 1,362, Visits: 15,269
chris.roddis-ferrari (11/15/2013)

Stats were updated for the MSreplication_subscriptions table.


I was wondering if you'd updated all the stats on the 2008 subscriber since upgrading the server?

I would consider running sp_updatestats on the db then look considering updating the stats with fullscan on the biggest tables at the subscriber.

Its possible that the different agent versions are using different execution plans for the replication stored procedures. Since the latency appears to be at the running of commands at the subscriber, this is where I would be looking.
Post #1514629
Posted Friday, November 15, 2013 3:11 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Friday, November 22, 2013 2:40 PM
Points: 16, Visits: 71
We had updated the stats on specific tables, but have now scheduled in an update with full scan for tomorrow morning
Post #1514644
« Prev Topic | Next Topic »

Add to briefcase ««123»»

Permissions Expand / Collapse