I don't use snapraid on windows but I do believe ReFs would be a good option for storing precisely the type of data snapraid users are interested in (mostly static, for a long time). It has (if enabled) checksums like ZFS.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
IMHO it would be desirable to support at least one modern file system on each platform with the ability to detect bit rot on every read. SnapRAID is a brilliant piece of software, but by design it can’t verify each and every single read an application performs on the data set.
Pretty much all of my hardware is server grade (Xeon CPU, ECC RAM, SAS controller, etc.). Still when I copy 1 TB of data between machines over the network I often get 1 or 2 bit flips. I don’t know how that’s possible, but I suspect that with some read patterns a disk may not always return the correct data and this goes undetected over the wire – despite my SnapRAID scrubs are always fine on the source machine. I think it is valuable to have an additional layer (the file system) verifying checksums every read.
Anyway, I’ll send another donation, not to push this request, just to say thanks for the continued support and dedication you put into SnapRAID!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Are there plans to support ReFS in the near future? Thanks.
I'll add it in the TODO list.
But I suppose not shortly. At now yours it's the first request.
Ciao,
Andrea
I don't use snapraid on windows but I do believe ReFs would be a good option for storing precisely the type of data snapraid users are interested in (mostly static, for a long time). It has (if enabled) checksums like ZFS.
IMHO it would be desirable to support at least one modern file system on each platform with the ability to detect bit rot on every read. SnapRAID is a brilliant piece of software, but by design it can’t verify each and every single read an application performs on the data set.
Pretty much all of my hardware is server grade (Xeon CPU, ECC RAM, SAS controller, etc.). Still when I copy 1 TB of data between machines over the network I often get 1 or 2 bit flips. I don’t know how that’s possible, but I suspect that with some read patterns a disk may not always return the correct data and this goes undetected over the wire – despite my SnapRAID scrubs are always fine on the source machine. I think it is valuable to have an additional layer (the file system) verifying checksums every read.
Anyway, I’ll send another donation, not to push this request, just to say thanks for the continued support and dedication you put into SnapRAID!