Pool configuration and format type suggestion

    exactly that was my concern about the zfs plugin, but I guess we will see, also thank you very much about the mergerfs explanation I dindn't know anything about it, so you saved me research :) it doesn't suit my needs.

    So onto the zfs then :)

    thank you Krisbee

    • Official Post

    mergerFS and SnapRAID are not well suited to a production environment. Note that there are additional (serious) risks in using mergerFS+SnapRAID that you may be unaware of:

    - Since SnapRAID relies on existing data for parity calculations, a large file delete after the last sync operation can affect the rebuild of a failed drive. Unrecoverable errors (lost or corrupted files on the rebuilt drive) would result.
    - Along similar lines, the data missing from a failed drive is unavailable until the rebuilt is complete. (This does nothing for user happiness.)
    - A drive rebuild is best done as an "off-line" process. If users are adding and deleting files during a rebuild process additional unrecoverable errors may result.
    - Restoration of folders and files is contingent on the last "sync" operation. As an example; if a file is deleted and a sync operation is completed before recovering the file, that file is beyond recovery. A sync resets the state of the SnapRAID array.

    This is why SnapRAID is not a good choice for the majority of business applications, particularly where a few to several users are involved. As has already been pointed out, it's only suitable for static data sets, like media libraries where files may be routinely added but file deletes are very rare.

    ___________________________________________

    Setting aside RAID0, none of the above are issues with any of the traditional RAID types (mdadm, ZFS, BTRFS, etc).

    you put in words exactly my experience crashtest , when I used snapraid back then I wasn't impressed, I had multiple errors on syncs and alike.

    Is always better to ask though as I used it many years ago, you never know, but yes I already ruled it out, was interesting especially the part on on mergerfs as I didn't know anything about it, I'm always happy to learn new stuff :)

    I will go full zfs experience then, hoping the server will not crap itself :D

    • Official Post
    Quote from Wek

    I will go full zfs experience then, hoping the server will not crap itself

    Will you have a backup server? It's possible to functionally replicate a duplicate of the main server and simply bring it on-line (or, more accurately, "make it visible" on the LAN).

    crashtest nope this would be overkill for the use at the moment.


    It's fine, they have plenty of offsite backup of the important data, so if the server will die for any reason they can keep working.

    In the future I'm planning to experiment with building another machine with proxmox as a base in that case I can think of making clusters and so on, but not for the moment :)

Participate now!

Don’t have an account yet? Register yourself now and be a part of our community!

Register Yourself Login