[rescue] New acquisition... (AIX)

Mike Meredith mike at blackhairy.demon.co.uk
Fri Apr 2 11:19:27 CST 2004


On Thu, 1 Apr 2004 17:17:11 -0600, Jonathan C. Patschke wrote:
> On the contrary.  I've been running it (in a somewhat decreased
> capacity since 2001) for 8 years.

Ah! A newbie then :) (I've been running it for 11-12 years).

The big trouble I have with your allegation that Linux is pretty much as
bad as Windows, is that I haven't seen that much trouble with Linux ...
if I thought it was as bad, I wouldn't still be running it on some of my
(less critical) servers ... they'd be running *BSD. 

> At the core, sure.  But there's a lot more than that core in the 50MB
> tarball that is the Linux kernel source.  

50 MBytes? Are you looking at 2.7.56 or something ? 2.6.4 is around
40Mbytes. A big chunk of that you'll never be running on any single
platform.

> There's a lot of useless
> crap and half-implemented drivers in there that I've been bitten by
> more than once.

Some of it is still under development. I've been bitten by driver
problems maybe once in 11-12 years, and as I recall the problem was
fixed in the next kernel release.

> the device naming scheme is brain damaged.  More than once I've
> upgraded a kernel on a machine with two different types of, say,
> Ethernet cards or SCSI cards, and.... Whooops!  eth0 is now eth1

I haven't experienced that myself, but I tend to have consistent
hardware ... same SCSI controllers, same NICs.

> because the device-probe stuff changed.  Every other Unix I've ever
> used (aside from AIX, which also does the lame enX naming) uses device
> names specifically to avoid this problem.

Actually they only half do it (which is probably good enough). If you
have two hme's in a Sun box, they're 'hme0' and 'hme1'.



More information about the rescue mailing list