Gluster users - Feb 2016 - Volume heal won't complete while I/O is running on v3.7.8

Hi all,



After replacing a brick in a disperse volume(2+1) with I/O running, I/O
works well but healing won't complete.


'gluster volume heal info' always shows that file with I/O running is
still
healing.


*Example:

*******************

Brick N9:/export/IFT_lvol_VOIVAqzlWd/fs

/_R at W_000

Number of entries: 1

*******************.


When I/O is stopped on the file after hours, healing completes after a few
seconds.

Is the 'file changing' cause the healing endless?


thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://www.gluster.org/pipermail/gluster-users/attachments/20160224/9f4d7dfa/attachment.html>

----- Original Message -----
> From: "???" <wowowowowody at gmail.com>
> To: gluster-users at gluster.org
> Sent: Wednesday, February 24, 2016 2:20:16 PM
> Subject: [Gluster-users] Volume heal won't complete while I/O is
running on	v3.7.8
> 
> 
> 
> Hi all,
> 
> 
> 
> After replacing a brick in a disperse volume(2+1) with I/O running, I/O
works
> well but healing won't complete.
> 
> 
> 
> 
> 'gluster volume heal info' always shows that file with I/O running
is still
> healing.
> 
Currently in EC there isn't a mechanism to detect the difference whether
files truly need heal or they *appear* to need heal due to the on-going I/O.
This is the reason why a file with I/O going on keeps showing up on heal info.

EC will provide a way to distinguish between these two scenario in
future.
> 
> 
> 
> 
> *Example:
> 
> 
> *******************
> 
> Brick N9:/export/IFT_lvol_VOIVAqzlWd/fs
> 
> /_R at W_000
> 
> Number of entries: 1
> 
> *******************.
> 
> 
> 
> 
> When I/O is stopped on the file after hours, healing completes after a few
> seconds.
> 
> Is the 'file changing' cause the healing endless?
> 
> 
> 
> 
> thanks
> 
> 
> 
> 
> 
> 
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-users
-- 
Thanks,
Anuradha.