Welcome to (posting on) Communities.
"When we design vSAN Hybrid scenario should we consider cache capacity for slack-space?"
No as no data is permanently stored here and thus they do not contribute capacity to the vsanDatastore.
"for example, I need 10TB usable capacity (not consider swap, snapshot to keep it simple), plan 30% of slack-space, FTT=1, the total requirement is (10TB*2)/0.7=28.6TB. According 10% cache general rule, how much cache should I plan? Is it 20TB*0.1 or 28.6TB*0.1?"
Actually in hybrid Cache-tier sizing is based on usable space e.g. what the usage is before FTT replicas - however do note that these recommendations are based on the fact that the average workload is accessing ~10% of its data at time/over a short period - if your workload varies greatly from this then bigger cache-tier devices may help improve performance.
"I think we don't need to consider cache for slack-space because it is a reserved space for maitain. Am I correct?"
Just to clarify, Cache-tier is divided up as 70/30% Read/Write cache, max usable write cache per device is 600GB at present.
The cache tier is just for caching, and does not count as "capacity". It will use 100% of the (up to) 600GB per cache drive, moving data in and out of the cache as needed.
Thank you for your response!
"It will use 100% of the (up to) 600GB per cache drive, moving data in and out of the cache as needed."
Actually for Hybrid (which OP is using) this is false - it will use (up to) 600GB for Write-cache if available and potentially more than this for Read-cache if available - it will always utilise this space in a 70/30 Read/Write ratio. Perhaps you are thinking of cache-tier in All-Flash which is 100% write buffer.
TheBobkin, you are correct. However, I was merely addressing the write buffer alone, as the read cache has no effect on the "slack space" subject. I was simply pointing out that the entire 600GB of the write buffer can be utilized without any need to consider any sort of "slace space". The only thing to really consider with an SSD when running at 100% utilized capacity is that write performance may suffer as a result of garbage collection, depending on how much over-provisioning the drive has.