Hi, so I was wondering about how you go about achieving redundancy in the write cache for the period between being written to the cache and the cache being flushed to the parity array. Do I need to configure two SSDs as a raid in the bios using Intel Rapid Storage, or do you just supply two disks and it takes care of it at software level in UNRAID?

 

Also I want to have a good amount of SSD space for VMs, plex transcoder cache and plex library. Do I need to buy a third SSD for that or can I use part of the mirror I either create in the first part of this question?

 

Finally, when I'm loading up the array with my content, it's pretty obvious that I'm going to run out of cache before it's flushed to disk at night, what happens in that instance?

The default configuration for the cache pool is btrfs raid1 so you get a mirror. It can be used as write cache, VMs, dockers, whatever you want. Other configurations are possible for the cache pool, and some people also use unassigned disks for VMs and dockers.

 

User shares set to cache-yes will overflow to the array. You need to be sure to set Minimum Free for cache (in Global Share Settings) to larger than the largest file you expect to write to cache. Unraid doesn't know how large a file will become when it begins to write it. If a disk has less than Minimum Free it will choose another disk to begin the write.

 

For the initial load, people often don't cache anything. Some even wait until the initial load is done before building parity so there is no write performance penalty for parity updates. There is also the Turbo Write setting to get better write speed with parity.

Some relevant links to the FAQ:

 

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/#comment-480421

 

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?page=2#comment-537383

 

 

 

Thanks for the prompt response. I'm going to buy 2 500GB Samsung 970 EVOs then!