[Gluster-devel] solutions for split brain situation

[Daniel Maher]

Stephan von Krawczynski wrote:
> Hello all,
> 
> we have seen several split brain situations and think that the most common
> option for the situation is simply missing. You can define a favourite child,
> but you cannot define to use the latest file copy as definitive. Why not?
> Isn't it a logical approach to say that the latest copy of a file based on
> mtime must be the most up-to-date and therefore being used in split brain
> recovery?

Are you sure about that ?  If you have one file which, due to split 
brain, now exists in two different states, and each of them are modified 
during the event, which one is actually correct ?

Imagine a log file ; the split brain event occurs, and now this log file 
exists in two different forms, both of which are being updated on either 
side of the split.  Each of these files contains data which, while 
incomplete, is still valid.  Clobbering the one with the older mtime 
(even, as in the case of log files, by perhaps seconds) means that 
you'll lose all of the log data that happened to be in the slightly 
older file.

-- 
Daniel Maher <dma+gluster at witbe.net>