What is the level of reliability of RAID0 under the condition of good hardware?

Question

RAID

For the MVP stage of the project, you need a large amount of memory for a couple of million zip archives, I want to raise a raid0 server from the following equipment:
2 14Tb SAS Western Digital (HGST) Ultrastar DC HC530 disks Broadcom
3008 controller built into the SuperMicro H11SSL-NC motherboard
critical and renewable and will not be accessed often.
Is my decision rational?
How to test an array before running to make sure that there will be no problems when filling the array?

Answer 1

What is the level of reliability of RAID0 under the condition of good hardware? - if there are two disks, then it is twice as bad as if there was only one disk.
Is my decision rational? - only if data loss doesn't bother you at all, I would look towards raid5
of 4x 8TB disks - there will be 24 TB of space instead of 28, but the array will withstand the failure of the 1st disk.
How to test an array before running to make sure that there will be no problems when filling the array? - do a load test run, and also check the array - filling it to 99%

Answer 2

Since raid0 dies when any of the disks fails, the probability of its failure is 2N, where N is the probability of one disk failure :)
Add to this the cost of data recovery (since they are renewable) and the replacement of the actual screw, and then think - is it necessary? Either way, it's up to you...

Answer 3

Is my decision rational?

Suitable for MVP, of course.
But instead of raid0, I would recommend having a separate file system for each disk, and then, if necessary, you can “blind” them using mergerfs. In this case, the death of one disk will lead to the loss of only the files stored on it.
Also, raid0 with a small block size is not very good in a multi-threaded load.