Xavi Hernandez

> > > > > > Do index insert/delete operation in an internal data structure(hash) and save the gfids in a file while index refcount is 1 and 0. During...

> > I have a concern from the security point of view: couldn't this feature leak information in cases it shouldn't ? I'm not sure what checks glusterd does to...

@amarts, is this PR replaced by #3745 ? if so, we should close this PR.

This is the high level roadmap I'm planning to execute to roll out the I/O framework globally for Gluster. Each step will require detailed analysis of the existing implementation to...

The analysis has been done using EC, but in theory the same problem exists with AFR. The issue could be avoided by sending some inodelk in DHT as internal operations...

Note about the order in which locks are acquired: The main GFID is always acquired first to make it backward compatible with other bricks/clients that may not be using the...

I've identified a problem when eager-locking is enabled (only in EC, AFR doesn't use eager-locking for non-regular files if I'm not wrong). The issue is that if the inode has...

After some thinking I've realized that the best way to fix this issue is actually to remove the dependencies between deletes and other operations on the inode being deleted. A...

You can run this command on any file of the volume: # getfattr -n glusterfs.pathinfo It will return (in a slightly scrambled way) the bricks and path where the file...

If you have sharding enabled then pretty much all files of a significant size (i.e. all VM disks) will have many shards on each subvolume. Why do you need to...