[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [ogfs-dev]The problems with resource group locking

On Thu, 2003-04-17 at 10:09, Cahill, Ben M wrote:
> Great info . . . will you add this to the locking doc?

I agree!  Nice job, Dominik!

> Also, what exactly is a "transaction", and "completing a transaction"?

A transaction is an indivisible operation that changes the state of
the filesystem.  A good example would be the creation of a new inode
which must perform multiple I/Os.  In this case, the directory must be
updated, an inode allocated, and the inode must be initialized with
the new contents.  The success of a create depends on the success of
all parts.  If the creation fails, all changes associated must also fail
(or be un-done) to not leave an inconsistent, or partially created file.

Transactions can be journaled (logged) to delay updating the actual
filesystem and to provide fast recovery in the event of a crash.  It's
faster because all known changes (transactions) are in the journal and
the entire filesystem does not have to be searched (as in fsck) to find
possible incomplete changes.  At the point when the journal entries are
written to the actual filesystem, they are considered "complete".

-- Joe DiMartino <joe@osdl.org>

This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
Opengfs-devel mailing list

[Kernel]     [Security]     [Bugtraq]     [Photo]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Clusters]     [Linux RAID]     [Yosemite Hiking]     [Linux Resources]

Powered by Linux