CCL might be whiny...
...but they're not wrong.
6 publicly visible posts • joined 3 May 2013
The article is about AFAs. All Flash Arrays, in this context meaning arrays designed from the ground up, which cannot take spinning disk.
Calling a compelllent lacking any spinning disk an all-flash array in this context is misleading. It's like calling an all-flash VNX an AFA.
You should check out the sla. The three nines is for service availability. If you get less than three nines you get some amount of free service. At no point do you get a refund.
Also, note the service availability fact. There is no SLA around data availability.
That sysadmin is full of shite. Firstly, any flipped bits on the raid set are going to be duplicated in the backup set. Second, all decent storage manufacturers maintain checksum data to detect flipped bits, and then use RAID to fix them. Which os why a failed disk in a raid-5 set is a dangerous place to be in if you don't have scrubber functionality in your array. Third, mature array technology often includes technology to do that aforementioned scrubbing. This dramatically reduces the incidence of a flipped bit causing a rebuild failure.