Groups | Blog | Home
all groups > sql server replication > september 2005 >

sql server replication : Log Reader Agent failing during cluster move group


DBA
9/14/2005 1:24:33 PM
All,
I am testing a clustering solution on Windows 2003 Enterprise and SQL
Server 2000 SP3a (Microsoft SQL Server 2000 - 8.00.760). Due to
unavailability of hardware we got 1 pyhsical (node1) and 1 virtual
machine (node2) as the 2 nodes of the cluster. SQL Server instance is
the remote distributor. The publication and subscribers are different
boxes and are not clustered.

The first test I performed is very simple that is to move the current
resources (including the Cluster Group, MSDTC Group and SQL Server
resource group) from the primary cluster node to the secondary cluster
node, and then back again. The steps are as follows:

Start Cluster Administrator.
Click on "Cluster Group" to highlight it. Note the "Owner" of the
resources, is node1 (physical machine)
First, right-click on "Cluster Group," then select "Move Group." The
"State" changed from "Online" to "Offline pending" to "Offline" to
"Online pending" to "Online." and the "Owner" changed from the primary
node to the secondary node that is from node1 to node2.
Did the same with MSDTC and SQL Server resource group with the same
result.

Everything came back up fine including all the replication agents (log
reader, distribution).

The problem arised when I did a move group from virtual to physical.
Everything worked as expected except the log reader agents. 1 log
reader agent gave an error "
The Step did not generate any output. The step failed.
"
and another log reader agent gave the error "
The process could not execute 'sp_repldone/sp_replcounters' on
'machineName\instancename'.
Another log reader is replicating the database.
(Source: machineName\instancename (Data source); Error number: 18752)
The process could not set the last distributed transaction.
(Source: 'machineName\instancename'-2 (Agent); Error number: 22017)
"

In both the cases stopping and starting the agent fixed the issue.

Has anyone faced this kind of issue? Is this a SQL Server issue or
clustering issue? Is this an issue due to Physical and Virtual machine
cluster or is this something that will occur even if we cluster 2
Physical machines (that is what we are going to do in production
environment)?

All help is appreciated.
Manoj
Hilary Cotter
9/14/2005 10:31:48 PM
I've seen this. I've also seen where the log reader agent will actually
attempt to write the same transaction to the distribution database.

Both events don't always occur, and I'm not sure how to fix them. As you
have noticed stopping the agent does seem to fix it.

--
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html

Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
[quoted text, click to view]

DBA
9/15/2005 9:58:28 AM
Does anyone know if this is a bug?
Is this a SQL Server issue or clustering issue?
Is this an issue due to Physical and Virtual machine cluster or is this
something that will occur even if we cluster 2 Physical machines ?
AddThis Social Bookmark Button