Configure Integration Services on a Cluster

We are attempting to setup Integration Services on a Cluster per the
following article: http://msdn2.microsoft.com/en-us/library/ms345193.aspx
We have the following Cluster Groups/resources
"Cluster Group"
Disk Q - Quorum
Disk R - MSDTC
MSDTC
MSDTC IP Address
Cluster Name
Cluster IP
"SQL1 Group"
Disk S - Data Files
Disk T - Log Files
SQL IP Address
SQL Network Name
SQL Server (SQL1)
SQL Server Agent (SQL1)
SQL Full Text (SQL1)
SQL1 Integration Services
"SQL2 Group"
Disk U - Data Files
Disk V- Log Files
SQL IP Address
SQL Network Name
SQL Server (SQL2)
SQL Server Agent (SQL2)
SQL Full Text (SQL2)
SQL2 Integration Services
"SQL3 Group"
Disk W - Data Files
Disk X- Log Files
SQL IP Address
SQL Network Name
SQL Server (SQL3)
SQL Server Agent (SQL3)
SQL Full Text (SQL3)
SQL3 Integration Services
The problem is that the integration services resources in the various
cluster groups won't come online on all nodes - only certain nodes. We're
getting an error indicating:
An error occured attempting to read properties for the 'xxx Integration
Services' resource.
The object already exists.
Error ID: 5010 (00001392)
Do you want to ignore this error and continue
[yes] [yes to all] [no]
Does anyone know how this should be setup? Should we have seperate
integration services resources in every group or only one group? Why are we
getting the error mentioned above?
Thanks,
Brad
You can only cluster it once. It is not an instanced component so the
service can only exist once on each node with the registry pointing at the
config file on the shared disk. Since you have multiple instances, follow
the instructions for "Configuring Integration Services in a Different Group
than SQL Server" to avoid dependencies on a specific instance being
available. If you store all your packages on the file system rather than
msdb this avoids any depence on any of the instances being available and you
can still share them out using clustered file shares to control access e.g.
Disk I
SSISPackages
SQL1 (Share)
SQL2 (Share)
SQL3 (Share)
SSISLogs
SQL1 (Share)
SQL2 (Share)
SQL3 (Share)
HTH,
Jasper Smith (SQL Server MVP)
http://www.sqldbatips.com
"Brad Baker" <brad@.nospam.nospam> wrote in message
news:eNXo%23ac7GHA.140@.TK2MSFTNGP05.phx.gbl...
> We are attempting to setup Integration Services on a Cluster per the
> following article: http://msdn2.microsoft.com/en-us/library/ms345193.aspx
> We have the following Cluster Groups/resources
> "Cluster Group"
> Disk Q - Quorum
> Disk R - MSDTC
> MSDTC
> MSDTC IP Address
> Cluster Name
> Cluster IP
> "SQL1 Group"
> Disk S - Data Files
> Disk T - Log Files
> SQL IP Address
> SQL Network Name
> SQL Server (SQL1)
> SQL Server Agent (SQL1)
> SQL Full Text (SQL1)
> SQL1 Integration Services
> "SQL2 Group"
> Disk U - Data Files
> Disk V- Log Files
> SQL IP Address
> SQL Network Name
> SQL Server (SQL2)
> SQL Server Agent (SQL2)
> SQL Full Text (SQL2)
> SQL2 Integration Services
> "SQL3 Group"
> Disk W - Data Files
> Disk X- Log Files
> SQL IP Address
> SQL Network Name
> SQL Server (SQL3)
> SQL Server Agent (SQL3)
> SQL Full Text (SQL3)
> SQL3 Integration Services
> The problem is that the integration services resources in the various
> cluster groups won't come online on all nodes - only certain nodes. We're
> getting an error indicating:
> An error occured attempting to read properties for the 'xxx Integration
> Services' resource.
> The object already exists.
> Error ID: 5010 (00001392)
> Do you want to ignore this error and continue
> [yes] [yes to all] [no]
> Does anyone know how this should be setup? Should we have seperate
> integration services resources in every group or only one group? Why are
> we getting the error mentioned above?
> Thanks,
> Brad
>
|||Would this work in the case of an active/active scenario where there
are 2 virtual SQL Servers, with one running on each of 2 cluster nodes?
When the integration packages are run, are they run by the integration
services service, or by the SQL Instance? In other words, does
Integration Services simply let you design the packages, or is it
running them?
I am finding this concept hard to grasp for some reason. I am used to
DTS in SQL 2000
I have a 2 node cluster with 2 virtual SQL servers, each with 1 named
instance. The Virtual servers can fail back and forth no problem. The
Integration Services is installed on both nodes, but gives an error
when I try and view the contents of the MSDB folder.
Jasper Smith wrote:[vbcol=seagreen]
> You can only cluster it once. It is not an instanced component so the
> service can only exist once on each node with the registry pointing at the
> config file on the shared disk. Since you have multiple instances, follow
> the instructions for "Configuring Integration Services in a Different Group
> than SQL Server" to avoid dependencies on a specific instance being
> available. If you store all your packages on the file system rather than
> msdb this avoids any depence on any of the instances being available and you
> can still share them out using clustered file shares to control access e.g.
> Disk I
> SSISPackages
> SQL1 (Share)
> SQL2 (Share)
> SQL3 (Share)
> SSISLogs
> SQL1 (Share)
> SQL2 (Share)
> SQL3 (Share)
> --
> HTH,
> Jasper Smith (SQL Server MVP)
> http://www.sqldbatips.com
>
> "Brad Baker" <brad@.nospam.nospam> wrote in message
> news:eNXo%23ac7GHA.140@.TK2MSFTNGP05.phx.gbl...