Hi,

The MaxConcurrentExecutables setting is set to -1. I guess that the decision what to run is not done by SSIS but the SQL server engine.

I have to take a deeper...

There is a top level task which drops all foreign key constraints. The tables are without any constraints (PK, Unique, FK, Check). Also, compression is disabled. Further, I use INSERT...

Cost Threshold of parallelism is set to 5 (I think it's default since stone age) and MAXDOP is number of logical cores of the machine (e.g. 10 physical cores with...

Hi,

thanks, that would be nice if you could come back to me with details. Links to literature is fine for me, I just need help to get the buzzwords that...

Hi Scott,

sorry for keeping you waiting so long. I had to made several tests before. I have read your link about minimal logging on clustered indices.

I have a very simple...

Oh, ok. I have to check that.

Is there any other option to avoid it? I have chosen to write all my staging data into an uncompressed heap due to minimal...

Thanks for the quick reply.

So, you think that MAXDOP as option will get the number of cores up for writing the table? Or is the writing always a single core...

Edit: I know of the following article:

https://docs.microsoft.com/en-us/sql/relational-databases/indexes/configure-parallel-index-operations?view=sql-server-ver15

But the machine is using all cores at the beginning. Does it switch the behaviour dynamically?

Thanks for the link. I do not use any of such operators. It is just a selection (simple copy of the source table), so no limitation from that side.

If I...

Ok, now I got you. I'll try it tomorrow when I am back at work :-). Hope that this will bring the benefit I was looking for.

Wow, never thought about that. But how does it work? If I write the SELECT statement in the SQL task and use a object variable as full return set how...

Hi,

thanks for links, the third one was new to me :-). It is indeed the nearest match to my use case. It still has a 1:1 relation between table and...

It is just a simple SELECT of a subset of columns. No transformation at all. I am loading my data from source system DBs to my staging area. All business...

I get your point to have full power over parallelism if I implement it manually.

But I don't get the part with the query processor. Why does the query optimizer allow...

I had table lock disabled in order to allow concurrent inserts. According to documentation minimal logging works for HEAP tables in SIMPLE RECOVERY and concurrent inserts....