[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gluster-users
Subject:    Re: [Gluster-users] Production cluster planning
From:       Lindsay Mathieson <lindsay.mathieson () gmail ! com>
Date:       2016-09-30 21:07:12
Message-ID: 07c25105-66fb-1aef-2916-b7eccab4df70 () gmail ! com
[Download RAW message or body]

On 1/10/2016 4:15 AM, mabi wrote:
> The data will not be in "any" state as you mention or please define 
> what you mean by "any". In the worst case you will just loose 5 
> seconds of data that's all as far as I understand.
>

By "Any" state I mean *Any*, you have no way of predicting how much data 
would be written.


The key thing here is that gluster will think the data has been safely 
written to disk when it has not. Imagine you have a three node volume 
undergoing heavy writes. Easily time for multiple shards or small files  
to be written in 5 seconds.


All of a sudden one of the nodes goes down - perhaps the cleaner 
unplugged it.


Now the third node is missing up to 5 seconds of data compared to the 
other two nodes - your gluster volume is in a inconsistent state and 
*doesn't* know it.

-- 
Lindsay Mathieson

_______________________________________________
Gluster-users mailing list
Gluster-users@gluster.org
http://www.gluster.org/mailman/listinfo/gluster-users

[prev in list] [next in list] [prev in thread] [next in thread]