Lessfs with journaling / checkpointing

Lessfs with journaling.

Lately I have been working on a new version of lessfs. The exciting new feature is that it does not need fsck after a crash. It’s can always recover to the last checkpoint / steady state.

Would you like to know more? Drop an email. ;-)

Example: mounting lessfs after a crash.

Feb 14 11:39:01 saturn lessfs[5698]: The selected data store is file_io.
Feb 14 11:39:01 saturn lessfs[5698]: Lessfs uses a 24 bytes long hash.
Feb 14 11:39:01 saturn lessfs[5698]: Automatic defragmentation is enabled.
Feb 14 11:39:01 saturn lessfs[5698]: cache 4096 data blocks
Feb 14 11:39:01 saturn lessfs[5698]: The Blue Midnight Wish hash has been selected.
Feb 14 11:39:02 saturn lessfs[5701]: Lessfs has not been unmounted cleanly.
Feb 14 11:39:02 saturn lessfs[5701]: Rollback to : Sun Feb 14 11:33:33 2010

Example : mounting lessfs after a clean umount.

Feb 14 11:53:51 saturn lessfs[5926]: The selected data store is file_io.
Feb 14 11:53:51 saturn lessfs[5926]: Lessfs uses a 24 bytes long hash.
Feb 14 11:53:51 saturn lessfs[5926]: Automatic defragmentation is enabled.
Feb 14 11:53:51 saturn lessfs[5926]: cache 4096 data blocks
Feb 14 11:53:51 saturn lessfs[5926]: The Blue Midnight Wish hash has been selected.
Feb 14 11:53:51 saturn lessfs[5929]: The filesystem is clean.
Feb 14 11:53:51 saturn lessfs[5929]: Last used at : Sun Feb 14 11:53:48 2010

This entry was posted in Uncategorized. Bookmark the permalink.

6 Responses to Lessfs with journaling / checkpointing

  1. Um, yes, very, very interested. :)

    Since LessFS is ideal for storing backups, reliability is important, and as I mentioned in an earlier comment, the fsck performance in my basic testing seems way too slow for real use.

  2. Alex says:

    Hi Mark,
    great news ! This could be so great ! As recovering crashes is such a pain for now !

    Let us know if this might work on the TC storage option ?

    This smell like such a revolution ;) hope to ear more about that stuff !

    By the way, but just as an idea for the far away future, getting replication, by, as an idea, while using the tc storage, storing datas on the first place and running the same data “updates” on a second place. Like your rollback but as a “rollnext” :)

    But first let’s know more about this new good news… :D

    • Mark says:

      Hi Alex,

      This option will work with both the tc and fileio backend.
      You can already implement replication with drbd. Lessfs will get native snapshots and replication in the future. DRBD does work and is available today.

      Mark

      • Mike says:

        Hi Mark,

        I’m not sure if it hit your spam or not, but I had sent an email, as you suggested, to inquire about how you’re avoiding the fsck.

        Any additional insight is greatly appreciated.

        Best,

        Mike

  3. bp says:

    Yes, having just started playing with lessfs, I’ve already had several lengthy lessfsck’s. Would love to see this feature!

  4. John says:

    Very interested.

    Question: Is Lessfs in its current form considered “production ready”?

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>