April 18, 2016 at 6:10 am
have a got at it and come back with any issues.
But as I said SSIS really isn't the best option for parallel processing like this.
in any case here is another implementation http://blog.exsilio.com/data-2/parallel-optimization-in-ssis-2012/
April 18, 2016 at 9:42 am
richardmgreen1 (4/18/2016)
@frederico - we did have a think about the load balancing option by using some sort of load table to pick up the next available file but couldn't quite figure it out. I'll see what I can do with your idea.@jeff - I didn't even think about the disks/spindles issue. I'm not sure where on the SAN our drives sit so that's another variable to put in the mix. Just a thought, if we could guarantee the use of SSDs on the SAN (I think we've got some) would that help?
If it's all on SSDs, there's no point in trying to do anything parallel because the SSDs won't come anywhere close to being the bottleneck.
--Jeff Moden
Change is inevitable... Change for the better is not.
April 19, 2016 at 1:57 am
Thanks for that Jeff.
I'll have a chat with the SAN admins and see where our drives are sat.
April 19, 2016 at 5:00 am
richardmgreen1 (4/19/2016)
Thanks for that Jeff.I'll have a chat with the SAN admins and see where our drives are sat.
While that will provide useful background info, this situation is crying out for comparative testing. I have had good results with parallelism in SSIS, despite having multiple targets on the same drive.
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
April 19, 2016 at 8:14 am
Phil Parkin (4/19/2016)
richardmgreen1 (4/19/2016)
Thanks for that Jeff.I'll have a chat with the SAN admins and see where our drives are sat.
While that will provide useful background info, this situation is crying out for comparative testing. I have had good results with parallelism in SSIS, despite having multiple targets on the same drive.
I absolutely agree. That's why I previously suggested that a 4 file test be done because "One good test is worth a thousand expert opinions".
--Jeff Moden
Change is inevitable... Change for the better is not.
Viewing 5 posts - 16 through 19 (of 19 total)
You must be logged in to reply to this topic. Login to reply