Author Topic: Suggestion for improvement to sRAID  (Read 6966 times)

Offline Benoire

  • Full Member
  • ***
  • Posts: 112
  • Karma: +0/-0
    • View Profile
Suggestion for improvement to sRAID
« on: January 10, 2016, 02:33:37 pm »
Morning Brahim,

While the title suggests an improvement, it is only one IF sRAID doesn't do this already.

When using the reFS filesystem, the filesystem is resistant to bitrot by its self-healing mechanisms however, when using Windows Storage Spaces the system can then find a perfect copy of the file by using the parity data to rebuild the live file if its checksum doesn't match.  This is a similar approach taken by BTRFS from what I can gather.

Is this something that could be put on a longer term development for sRAID?  As Microsoft improve reFS it should become a decent enough filesystem for normal use and having it not only self-heal, but also restore from parity for a failed checksum would be fantastic; Would definitely add value I think.

I guess thinking about it, this could also apply to any of the flexraid family that is using reFS formatted volumes as long as parity is stored.

Chris

Online Brahim

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 8,485
  • Karma: +201/-16
    • View Profile
Re: Suggestion for improvement to sRAID
« Reply #1 on: January 10, 2016, 03:47:46 pm »
This feature will part of full NZFS stack - a discussion for much, much, later. ;)

Offline Benoire

  • Full Member
  • ***
  • Posts: 112
  • Karma: +0/-0
    • View Profile
Re: Suggestion for improvement to sRAID
« Reply #2 on: January 10, 2016, 04:18:47 pm »
I did wonder given your hints on the NZFS stack but as we're only seeing snippets of what it can do I thought I'd post it here.  Are you making NZFS cluster aware as that is one of the biggest problems with Windows filesystems to date, unless you're on big dollars for IBM/Oracle etc.

Is it worth creating a suggestions part of the forum for your products where people can post ideas for enhancements that you could look at?