[Gluster-users] Ext4 considered harmful (Was: Re: Slow read performance)

Jeff Darcy jdarcy at redhat.com
Mon Mar 11 21:51:25 UTC 2013


On 03/11/2013 03:07 PM, John Mark Walker wrote:
> It didn't affect distributions for quite a few months afterwards. Having said
> that, your point is valid. We should have been more proactive on this front.


I'm inclined to agree.  On the one hand, I think there are some people who just 
don't understand just *how* disruptive this change was, how contrary to both 
the "don't break userspace" rule and basic software-engineering discipline.  On 
the other hand, eight months is an awfully long time for a patch to remain in 
review.  ("A year" is exaggerating a little, because of distro packaging etc.) 
  It's quite reasonable to consider all of the above to be insufficient excuse 
for such a long delay.  Ditto for other demands on our time, XFS as an 
alternative (which gave us some other problems at exactly the same time BTW), 
and so on.  It should have been fixed by now.

I wish I had something better to say than "sorry" and "we'll try to do better" 
but I'm afraid I really don't.  Let's try to get this into 3.4beta1 and put it 
behind us.



More information about the Gluster-users mailing list