Rich Freeman on 20 May 2015 12:27:24 -0700

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: [PLUG] Ext4

On Wed, May 20, 2015 at 9:03 AM, Anthony Martin
<> wrote:
> And people always questioned why I don't use raid on my home systems lol.

I think the bigger issue here is QA in general, and less so the details.

People point to Heartbleed and say "what if nobody compiled in that
heartbeat function that nobody uses anyway?"

Anytime there is a bug that is triggered by specific circumstances
people are going to argue "if only."  The problem is that when the
next bug comes along which is mitigated by those circumstances then
you don't hear them chiming in so loudly.  :)

That said, there is something to be said for avoiding exotic
configurations (says the guy who runs Gentoo).  The closer you are to
mainstream, the more likely you are going to be covered by various QA
systems that discover problems before you do.

However, ext4 on mdadm is hardly an exotic configuration.  A big
problem here is that despite the whole many-eyes argument, I don't
think there are really that many organizations doing serious
regression testing of key kernel facilities like filesystems.  Bugs
creep into even stable branches of the kernel, as I've experienced
recently with btrfs, and as we're seeing here with ext4.

You're probably safer with something like RHEL with its very long-term
kernels, as long as you don't expect it to work with any piece of
hardware made in the last two years and you stick to cutting-edge
filesystems like ext3.  :)

Philadelphia Linux Users Group         --
Announcements -
General Discussion  --