Does anyone have a recommendation for the processes distributed between the two machines? Background? Repository? Thanks
1 of 1 people found this helpful
Hi James -
I suspect the reason you haven 't gotten a reply is because the answer really is "it depends", and folks (myself included) don't want to give you a one-size-fits all answer that doesn't fit your particular situation. There are a number of different variables in play, and it'd be pretty difficult to work through everything via a forum posting.
That being said, here are some thoughts. I think you'll probably need to experiment a little to find the settings that work best for you, however.
- 4 VizQLs on a machine that only has 16 GB of RAM sounds a bit too much to me - The VizQLs could consume all the RAM on the machine, without leaving anything left for the OS or other Tableau processes. Too many VizQLs will also bring down your Cache Hit ratio which results in lower "perceived performance" ffor your users. Consider lowering this to 2 unless you have a specific reason why set it at 4.
- I like the idea of putting your backgrounders on a distinct machine. The Data Engine is also resource intensive process, so you might might want to consider putting it on the same box as the Backgrounders, too. Since your extract refreshing work will happen at night, that leaves the Data Engine with lots of room during the day to do it's work.
Hope this helps?
Thank you Russell,
I think we'll split the VizQL to 2 and 2 between the machines, drop the Data Engine to the worker, and put our background processes on the worker. I think we'll start with 4.
I appreciate it,
BTW - I meant we are on a 64 bit system.