Custom Task Deployment

Hello All,

We have developed some custom tasks which we currently deploy the standard way. That is, we install the custom task assemblies in the DTS\Tasks folder and we also install the assemblies in the GAC. Due to some deployment practices we are trying to implement we would like to be able to remove the assemblies from the GAC and install them in some other location. I have tried installing the assemblies in the DTS\Binn folder, but this does not appear to be working. When I drop the custom control flow tasks onto the package's control flow tab I get the "Cannot create a task with the name ..." error.

Is it possible to not have the custom task assemblies in the GAC? Are there any tricks to getting it to work?

FYI, our assemblies are signed. I'm not sure if that has anything to do with our troubles.

Thanks,

Rob

1234123412 wrote:

Hello All,

We have developed some custom tasks which we currently deploy the standard way. That is, we install the custom task assemblies in the DTS\Tasks folder and we also install the assemblies in the GAC. Due to some deployment practices we are trying to implement we would like to be able to remove the assemblies from the GAC and install them in some other location. I have tried installing the assemblies in the DTS\Binn folder, but this does not appear to be working. When I drop the custom control flow tasks onto the package's control flow tab I get the "Cannot create a task with the name ..." error.

Is it possible to not have the custom task assemblies in the GAC? Are there any tricks to getting it to work?

FYI, our assemblies are signed. I'm not sure if that has anything to do with our troubles.

Thanks,

Rob

Unfortunately (for you) they are required to be in the GAC.

-Jamie

|||

It makes sense (standard .Net assembly loader stuff), and as indicated in the post below , the assemblies can be loaded from the execution host directory as well. It is not documented, and probably not supported either. I know that the designer uses the DTS\ObjectType folders, so have you tried it in DTS\Task and DTS\Binn at the same time? If so and it still does not work, then why not try the supported/documented method ;) In theory the designer only requires the DLL in the DTS\Task folder, the GAC is used at runtime.

The only other reason for the error, is that the assembly you are trying to add, as stored in the toolbox "metadata" is no longer the assembly you actually have. Maybe clearing out the toolbox will help. If you changed the strong name, then you will have to fix up any packages and also sort out the toolbox.

Re: Custom SSIS Task Deployment - MSDN Forums
(http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=803910&SiteID=1)

Why not install them in the GAC as usual?

|||

Our issue is that we have several data access layer assemblies which are shared by our custom tasks and some web applications. The problem is that on several of our developer machines we are having build problems when building the web applications. We get error messages when we try to register the assemblies in the GAC. It has something to do with the ASP.Net worker process "locking" the assemblies even though the web application isn't running.

I haven't been able to get the custom tasks to work if they aren't in the GAC, but since the web application/GAC issue is just with the DAL assemblies it is really the DAL assemblies that we want to avoid putting in the GAC. I am able to put the DAL assemblies in the DTS\binn directory and keep the custom task assemblies in the GAC. This solves our problem.

Thanks for your repsonses.

custom path for data files

Hi,
I am about to install sql 2005 in a clustered scenario. I am reading the
installation notes and it says :
On the Cluster Group Selection page, select the group that contains the
drive you want to use for SQL Server from the Available Cluster Groups. The
selected cluster group is where the SQL Server virtual SQL Server resources
are placed. If you select the group containing the cluster quorum resource, a
warning is displayed recommending you do not install to the cluster quorum
resource. You can also specify a custom path for data files in the Data Files
field. To proceed, click Next.
My question is that I have two nodes both sharing D, E & QDrive if I select
the custom path option at this point in the installaion will I be able to
select D for databases and E for logs?
http://msdn2.microsoft.com/en-us/library/ms179530.aspx
Yes. You can add E: to the cluster group, make SQL dependent on it, and use
it for data or logs.
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP
"Paul" <Paul@.discussions.microsoft.com> wrote in message
news:99A9574F-1DDF-4741-8676-FC6EB36DFAFB@.microsoft.com...
> Hi,
> I am about to install sql 2005 in a clustered scenario. I am reading the
> installation notes and it says :
> On the Cluster Group Selection page, select the group that contains the
> drive you want to use for SQL Server from the Available Cluster Groups.
> The
> selected cluster group is where the SQL Server virtual SQL Server
> resources
> are placed. If you select the group containing the cluster quorum
> resource, a
> warning is displayed recommending you do not install to the cluster quorum
> resource. You can also specify a custom path for data files in the Data
> Files
> field. To proceed, click Next.
> My question is that I have two nodes both sharing D, E & QDrive if I
> select
> the custom path option at this point in the installaion will I be able to
> select D for databases and E for logs?
>
> http://msdn2.microsoft.com/en-us/library/ms179530.aspx
|||Hi Geoff,
Thanks for this but I'm new at this and a little confused.
I have:
cluster group containing Q
group 0 containing D Drive (I would like the data files here)
group 1 containing E Drive (I would like the log files here)
When I went through the setup I was expecting to have the choice of
splitting the files but could only chose one cluster group so I chose cluster
group 0.
Now when I opened management studio to change the logs location I can only
see D, adding the E Drive to group 0 also does not allow me to see E drive
from management studio.
What am I missing?
thanks in advance
I have now installed sql and at the point when asked
"Geoff N. Hiten" wrote:

> Yes. You can add E: to the cluster group, make SQL dependent on it, and use
> it for data or logs.
> --
> Geoff N. Hiten
> Senior SQL Infrastructure Consultant
> Microsoft SQL Server MVP
>
> "Paul" <Paul@.discussions.microsoft.com> wrote in message
> news:99A9574F-1DDF-4741-8676-FC6EB36DFAFB@.microsoft.com...
>
|||You are missing dependency.
Using the Cluster admin tool, take the SQL Service offline but leave the
rest of the group online. Right-click the SQL Service... Properties.
Advanced tab. Add the new disk as an item the SQL Service is dependent on.
Bring SQL back online.
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP
"Paul" <Paul@.discussions.microsoft.com> wrote in message
news:1AB63836-C0CF-4228-A1DC-9D8F28AB63FA@.microsoft.com...[vbcol=seagreen]
> Hi Geoff,
> Thanks for this but I'm new at this and a little confused.
> I have:
> cluster group containing Q
> group 0 containing D Drive (I would like the data files here)
> group 1 containing E Drive (I would like the log files here)
> When I went through the setup I was expecting to have the choice of
> splitting the files but could only chose one cluster group so I chose
> cluster
> group 0.
> Now when I opened management studio to change the logs location I can only
> see D, adding the E Drive to group 0 also does not allow me to see E drive
> from management studio.
> What am I missing?
> thanks in advance
>
> I have now installed sql and at the point when asked
> "Geoff N. Hiten" wrote:

Custom Install

I'm installing SQL for the first time in my Lab and I'm using the
Microsoft SQL 2000 Training kit. During the custom install I have the
option to install the Program files and the Data files in separate
partitions however no where in the book does it suggest that either
partition would be the optimal location for performance reasons. It
seems to me that separating the program files from the data files
makes more sense. If I had a production server with 200 or 300GB in a
Raid array I don't think I would make one big system partition and
install the whole thing on it. But thats why I'm asking, what say you
all? What do the Pros do?
Tia
ScottThe important thing is where you later create your database files for your production databases. The
program files are largely irrelevant as are the system database's files (except perhaps tempdb). In
general, put the transaction log file(s) on its own RAID1 and the database on RAID 10, 01 or
possibly 5.
There are plenty written on that topic. One place you could start is: www.sql-server-performance.com
--
Tibor Karaszi, SQL Server MVP
Archive at: http://groups.google.com/groups?oi=djq&as_ugroup=microsoft.public.sqlserver
"scott abra" <scott_abra@.hotmail.com> wrote in message
news:44e8e747.0310231516.41503a97@.posting.google.com...
> I'm installing SQL for the first time in my Lab and I'm using the
> Microsoft SQL 2000 Training kit. During the custom install I have the
> option to install the Program files and the Data files in separate
> partitions however no where in the book does it suggest that either
> partition would be the optimal location for performance reasons. It
> seems to me that separating the program files from the data files
> makes more sense. If I had a production server with 200 or 300GB in a
> Raid array I don't think I would make one big system partition and
> install the whole thing on it. But thats why I'm asking, what say you
> all? What do the Pros do?
> Tia
> Scott