[Linux-cluster] locking granularity of GFS?

[Michael Conrad Tadpol Tilstra]

On Thu, Jun 02, 2005 at 05:34:44PM -0400, Dan B. Phung wrote:
> What is the current locking granularity of GFS?  I found an IBM GPFS
> report that stated the GFS uses a global lock, which is entirely false
> since I know GFS uses DLM.  
> 
> From what I've seen from use, my guess is that it's at the directory
> level, which seems wierd.  Writing to different files in the same
> directory gives sequential finishing times.  when I modified the output to
> go to different directories, this change alleviated the finishing times
> disparity.

GFS has a couple pf locks for each file. (one for data, one for meta
data, one for iopen counts. maybe others, don't recall off the top of my
head.)  Directories get a lock, as well as most of the interal
structures.  So more-or-less gfs locks at the file level.  (note that
this is not the same or similar to fcntl locking, nor is it compatible.)

-- 
Michael Conrad Tadpol Tilstra
I can read your mind, and you should be ashamed of yourself.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20050603/4102c408/attachment.sig>