Re: Debian and Kernel Releases
In article <cistron.Pine.LNX.3.96.990320124609.7000B-100000@calvin.captech.com>,
George Bonser <grep@oriole.sbay.org> wrote:
>On 20 Mar 1999, Miquel van Smoorenburg wrote:
>
>> If you follow the kernel mailing list you'll see that even the current
>> 2.2.3 kernel still has quirks. On some machines here I can't even run
>> it, the NFS server keeps dying. 2.0.36 is rock-stable.
>
>That is only if you use the new kernel NFS. With user NFS (the same type
>zs 2.0.36 uses) it is stable as well.
Not on our servers. With a slink user-level NFS server and a 2.2.2 kernel
the NFS server kept dying every 2 hours or so. Went back to 2.0.37-preX
and it's rock stable now. I haven't hade the time to debug this yet,
real work got in the way ..
Another thing, there are still bugs in the routing code. Alexy has
had patches for this since 2.1.11x but they never went into the
kernel. They only just appeared in the latest -ac patch. There
are some more things like this, and it just takes a bit of time to
get all these issues worked out.
>And even then 2.2.x works fine on
>the VAST majority of systems. There might be one or two out there but
>overall 2.2 runs on a lot more systems than 2.0 does. 2.1 was the
>development kernel, 2.2 is a stable release. It is a lot more stable than
>2.0 was when it was released. We went to 2.0.6 in what, a week or two?
Yes. Still, I claim 2.2.x isn't stable enough yet for production use
in a business environment with a lot of networking just yet. Let it
level out a bit. Getting 2.2.latest into potato which will be released
in 2 or 3 months (we hope) is a reasonable goal though. By then 2.2.x
will be stable enough. Ofcourse this is exactly what Debian is doing,
I should have pointed that out in my earlier message.
Mike.
--
Indifference will certainly be the downfall of mankind, but who cares?
Reply to: