[Gluster-users] gluster heal vs gluster heal full

[Vijay Bellur]

On 11/28/2013 07:27 PM, Diep Pham Van wrote:
> On Thu, 28 Nov 2013 13:10:00 +0100
> "Maik Kulbe" <info at linux-web-development.de> wrote:
>
>> is there a difference between the command `gluster volume heal <vol>`
>> and `gluster volume heal <vol> full`?
>
> gluster volume heal <vol>: heal files that needs healing

gluster volume heal <vol> heals files that are part of the index built 
by self-heal daemon. The index refers to contents in 
<brick>/.glusterfs/indices/xattrop/ directory.

> gluster volume heal <vol> full: heal all files

A "full" self heal crawls through the entire volume and looks for 
possible candidates that need to be healed. The preferred method is to 
go through the index based heal and "full" heal is typically designed to 
be used in scenarios where there has been a disk/brick replacement or a 
re-formatted brick etc.

>> And is it possible to cancel a
>>> heal once started?
> It is impossible.

self-heals initiated by self-heal-daemon can be cancelled by killing the 
self-heal-daemon or through this command:

volume set <volname> self-heal-daemon off

-Vijay