disParity » General Discussion

New User questions

(9 posts)
  1. Sparks
    Member

    I did a create on a small sample of data and followed with a hashcheck and then a verfiy and found the hashcheck was faster. What's the verify doing if not a hashcheck on each file?

    Also, would you be able to rename the prefix the log files with the function that it's reporting...such as Create log 09-12-23 16.27.45.txt.

    Posted 5 years ago #
  2. Roland
    Roland

    Verify checks that every block of parity data matches the file data that it was calculated from. It's a lengthy operation since every single parity block must be recomputed (similar in time to a create.) It's main purpose is to detect bugs in disParity itself, since a bug is in theory the only reason a verify would fail.

    Hashcheck doesn't check the parity data, or even look at it. It recomputes the MD5 hash of every file on the drive and checks it against disParity's stored hash for that file. It's main purpose is to detect file corruption (which I've heard is becoming more likely with today's multi-TB hard drives, although I've never personally seen it happen.) If a file has become corrupted since the last update, the hash values won't match and disParity will report this.

    Posted 5 years ago #
  3. Sparks
    Member

    When I run a verify....it looks like it does a quick check to see if something has changed (Adds, Moves, Deletes) before going into it's block check routine. What's it doing there?

    Going back to the hashcheck question.....so it's just comparing a value already stored. Even if the hashcheck value is correct.....the parity could be corrupt. The hashcheck is validating the original file isn't corrupt or hasn't been changed.

    Posted 5 years ago #
  4. Roland
    Roland

    Yes, the first thing verify does is check whether anything has changed on the drive (the same process done in setting up for an update) because any changes would cause the verify to fail, since the parity is now out of date.

    Yes, hashcheck only checks the files, not the parity, so if the parity was corrupt hashcheck would not detect it.

    Posted 5 years ago #
  5. Sparks
    Member

    Would it be safe to say that when you run a Verify and if it got past the initial check with 0-Adds,0-Moves, 0-Deletes,0-Errors, then the data should match the parity? If so, would you be able to create a new command just to use the first part of a Verify to have a quick check if anything changed before doing a full block check Verify...maybe call it Pre-Verify. The reason I say this is that if you do a Verify before an Update after you have Added a file....it contiunes on with the block check portion of the Verify....or is this a bug? If that is the case...guess that's what it's suppose to do anyway. I guess you could always do an Update before a Verify to make sure you are catching all changes....but I'm not sure how long that takes since I'm only using a small sample of data.

    Hope this all makes sense.

    Posted 5 years ago #
  6. Roland
    Roland

    Yes, the expectation is that if a verify runs to completion (no deletes or edits detected) then parity should match the data. This initial check is really just to avoid having an expensive, time-consuming verify fail halfway through because some file got deleted or moved.

    An added file doesn't cause the verify to abort, because files added since the last update won't cause a verification error (verify checks the parity data against the files that were present at the time of the last update, so files added since then aren't in the parity and won't be checked.)

    When I said "any changes" in my post above I really should have said any deletes, edits, or moves.

    Posted 5 years ago #
  7. Sparks
    Member

    Sounds good.

    What do you think of the idea of renaming the log files...such as Create log 09-12-23 16.27.45.txt?

    Posted 5 years ago #
  8. Roland
    Roland

    I guess I'm not sure I see the purpose. I like the current naming scheme because it sorts alphabetically by date and time. I could possibly see putting the command at the end of the name (after the time and date) but not at the beginning.

    Posted 5 years ago #
  9. Sparks
    Member

    I see your point about sorting. I'm not sure it's a big thing...but I think it would add some value. If you could add it to the end without much trouble....that would be great.

    Posted 5 years ago #

RSS feed for this topic

Reply

You must log in to post.