VNX
Posts
Unsolved
This post is more than 5 years old
31 Posts
0
919
September 10th, 2013 09:00
MultiCore FAST Cache
How does MultiCore FAST Cache on a VNX keep one LUN from hogging all of the SSD space?
Responses(4)
No Events found!
Unsolved
This post is more than 5 years old
31 Posts
0
919
How does MultiCore FAST Cache on a VNX keep one LUN from hogging all of the SSD space?
Top
ScottPhelps
31 Posts
2
September 10th, 2013 09:00
By default 50% of the FAST Cache capacity is spread equally among all LUNs ensuring that every LUN gets some space in SSD for it's busiest data.
The other half of the FAST Cache SSD space is reserved for LUNs to use on an as need basis.
I think of it much like a private and a public space. The private space is a little piece that is carved out for every individual LUN to use, but the public space is used by the busiest 64K chunks in the busiest LUNs as needed based on which chunks are the "hottest".
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
September 10th, 2013 10:00
any new granularity, can i disable Fast Cache for individual pool LUN ? Any additional improvements in terms of detecting Fast Cache "un-friendly" workload (sequential) and not promoting it ?
ScottPhelps
31 Posts
0
September 10th, 2013 10:00
Roger that...if it hits SSD in a FAST VP pool, no need to move it to the SSD in FAST Cache. That SSD cache space is reserved for the "stuff" that still lives on spinning media.
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
September 10th, 2013 10:00
same rules apply where if block is already in FAST-VP tier (flash), it will not be promoted ?