15 Jul
2004
15 Jul
'04
9:43 a.m.
I understand this - however when the kernel developers are taking an interest
in the problems with audio/latency in the current kernel and are actively
asking for good bug reporting then in an effort to improve things it is
probably better if more people started testing the 2.6 branch.
Dave
On Thursday 15 July 2004 01:44, Malcolm Baldridge wrote:
People, people...
Speaking as someone who's used and tinkered with Linux from the 1.x days,
never ever be fooled into thinking that "release" Linux kernels mean
release as in "high quality and/or bug-free" (as bug-free as can be
realistically expected at least).
I avoid new kernels until they make it through adolescence... .17 or .18.
Seriously. I could regale you with sordid tales of mid-stream VMM changes,
driver breakage, filesystem miseries (for ext3!!), etc. But just take my
word for it.
Unless you want to be on the BLEEDING edge (make sure you remember the
operative word there), just stick to the mature kernels. 2.4.26 is nice
and solid, and whatever flaws remain with it are really minor ones. I use
for everything from SMP web/MySQL servers, to
everything-including-the-kitchen-sink home network servers, to personal
development stations. It works, and it works well. There are mature and
well-tested low-latency patches which just work well. I use them on my
production servers and notice improvements in many aspects of their
performance, without compromises in uptime or robustness. (Many of these
servers crossed over a year of uptime recently.)
Let those who have an ample supply of band-aids and H2O2 mess around with
2.6. They're still getting their infrastructure issues sorted out.
=MB=