[GFS2] journal recovery patch
authorDavid Teigland <teigland@redhat.com>
Thu, 20 Apr 2006 21:03:48 +0000 (17:03 -0400)
committerSteven Whitehouse <swhiteho@redhat.com>
Thu, 20 Apr 2006 21:03:48 +0000 (17:03 -0400)
commitc63e31c2cc1ec67372920b5e1aff8204d04dd172
tree950b2537b5a7e83e35a14a973da7bac84e844a52
parent190562bd84a484bf6590425aa2bb4d6d611c112b
[GFS2] journal recovery patch

This is one of the changes related to journal recovery I mentioned a
couple weeks ago.  We can get into a situation where there are only
readonly nodes currently mounting the fs, but there are journals that need
to be recovered.  Since the readonly nodes can't recover journals, the
next rw mounter needs to go through and check all journals and recover any
that are dirty (i.e. what the first node to mount the fs does).  This rw
mounter needs to skip the journals held by the existing readonly nodes.
Skipping those journals amounts to using the TRY flag on the journal locks
so acquiring the lock of a journal held by a readonly node will fail
instead of blocking indefinately.

Signed-off-by: David Teigland <teigland@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
fs/gfs2/ops_fstype.c
fs/gfs2/recovery.c
fs/gfs2/recovery.h