[Linux-cluster] GFS2 development plans

[Steven Whitehouse]

Hi,

On 13/11/15 08:13, Milos Jakubicek wrote:
> Hi,
>
> can somebody from the devel team at RedHat share some thoughts on what 
> are actually the development plans for GFS2 in the future?
>
> I mean: will it mainly be small and big performance improvements like 
> in the past couple of years, or are there any major new things 
> planned? In particular, I would be interested to hear about online 
> fsck -- whether this is something that is at least being considered 
> (on really big arrays with a large number of partitions around 10TB, 
> fsck is really a pain).
>
> Thank you
> Milos

Well the best way to see what is currently going on, is to watch the 
cluster-devel at redhat.com list, where you can see all the latest 
discussions. Since this is open source, it is not possible to give a 
definitive answer to that question - at any time someone might come 
along with a new feature upstream that we were not expecting!

 From a purely Red Hat PoV, we try to listen carefully to what our 
customers & community are asking for when we decide what to work on 
next. So if there is something specific that you'd like to see, then 
we'd definitely like to know what that is, and (perhaps even more 
importantly) what the use case is.

Better error recovery is something that we'd like to do. It is unlikely 
to be possible to implement all the checks that fsck does online, but I 
hope that it will become increasingly possible to recover from certain 
kinds of error without requiring fsck for the very reason that you've 
mentioned above.

Overall our main priority is to maintain stability, and this has to 
override all other considerations, so any new features will have to go 
through a lot of testing, and we work to improve the testing just as 
much as the filesystem itself,

Steve.