glusterfs.git/xlators/cluster/afr/src/afr-inode-write.c, branch v2.0.4

Do not lock entire file for metadata operations.

2009-05-05T06:41:54+00:00

Metadata operations now lock only a single byte at offset
(LLONG_MAX - 1) instead of the whole file, to avoid contention
with writev().

Signed-off-by: Anand V. Avati

Cleaned up log messages in replicate.

2009-04-24T15:30:14+00:00

Signed-off-by: Anand V. Avati

Changed xattr format of afr changelog to support adding and removing of subvolumes while keeping existing data.

2009-04-16T09:11:41+00:00

Signed-off-by: Anand V. Avati

update cluster/afr with new readv writev prototypes

2009-04-12T06:03:47+00:00

Signed-off-by: Anand V. Avati

Fix in changelog logic.

2009-04-07T17:49:08+00:00

If a writev fails, remember it by marking it in the fd context.

Signed-off-by: Anand V. Avati

Consider a subvolume dead if an fop fails on it

2009-04-07T17:49:00+00:00

Transaction fops earlier called afr_transaction_child_died only
if an fop failed due to ENOTCONN or EBADFD. Now they consider a child
dead regardless of the reason for failure. This handles cases such
as ENOSPC.

Signed-off-by: Anand V. Avati

2009-02-26T16:48:46+00:00

updated copyright header to include 2009.

Signed-off-by: Anand V. Avati

Added all files

2009-02-18T12:06:07+00:00