Very Computer

by **Robert Hyat** » Thu, 21 May 1998 04:00:00

Just a short note... I've been running 2.1.102 since it came out
last week, and, to date, have had no panics, no oopses, no ughs,
no aiee!!'s nor anything else untoward, running with 4 pentium
pro cpus, reading writing big disk files, lots of net traffic,
PPP connections, Xwindows, and all the other assorted stuff one
does with Linux.

Looks good... In fact, the last dozen versions or so have been
rock solid for me, at least starting with 2.1.97 and working
forward, although 2.1.91 and later looked pretty good.

Looks like progress... waiting on Linus's new "looks so good it
must be right" interrupt handler and version 2.1.103 now. :)

--
Robert Hyatt Computer and Information Sciences

(205) 934-2213 115A Campbell Hall, UAB Station
(205) 934-5473 FAX Birmingham, AL 35294-1170

by **Wayne Hyd** » Thu, 21 May 1998 04:00:00

> Just a short note... I've been running 2.1.102 since it came out
> last week, and, to date, have had no panics, no oopses, no ughs,
> no aiee!!'s nor anything else untoward, running with 4 pentium
> pro cpus, reading writing big disk files, lots of net traffic,
> PPP connections, Xwindows, and all the other assorted stuff one
> does with Linux.
> Looks good... In fact, the last dozen versions or so have been
> rock solid for me, at least starting with 2.1.97 and working
> forward, although 2.1.91 and later looked pretty good.
> Looks like progress... waiting on Linus's new "looks so good it
> must be right" interrupt handler and version 2.1.103 now. :)

I haven't bothered updating to .102 as .101 is working great for me.
(Hopefully there isn't an important reason to do so) My uptime is
currently 10+ days, and I have been trying to make it fall over without
success. The system is a dual PPro, btw, so SMP works fine for me. I
did have problems with 2.1.99, but it may have been due to the software
RAID. I've since split up the scsi disks to a non-RAID config. (seven
HD's, two SyJet's, CD, and tape over three BusLogic controllers -- two
958, one 948)

Has anyone had any problems with the md driver on recent 2.1 kernels? I
got a bunch of solid freezes and a few "Inactive in
scsi_request_queuable" messages sent to the console. I don't know if it
actually was the md driver, but I'm working fine without it now. I
upgraded to .101 from .99, so that might have resolved the problem and
the md driver was really safe in the first place. (I was using RAID0
with a good tape backup schedule.)

For the first five days my load average was about 5.5 with a few large
compiles being done over and over, etc. Since then I've toned it down
to about 2.0 to 3.0 with rc5 cracks and mp3 encoding and a bunch of
network activity. The Linux box is my http/ftp (squid) cache, handles
my ppp connection and masqurading -- it has two network cards (one 10Mb
to a SPARC4 and a 100Mb to a hub w/ a Win95 box and sometimes others).
Pretty sad when I'm using a SPARC4 only as an xterm into my PC Linux
box....

The only thing I haven't stressed is the sound. I've pretty much been
playing mp3's on the Win95 box.

-Wayne

by **Sid Boyc** » Sun, 24 May 1998 04:00:00

> Just a short note... I've been running 2.1.102 since it came out
> last week, and, to date, have had no panics, no oopses, no ughs,
> no aiee!!'s nor anything else untoward, running with 4 pentium
> pro cpus, reading writing big disk files, lots of net traffic,
> PPP connections, Xwindows, and all the other assorted stuff one
> does with Linux.

> Looks good... In fact, the last dozen versions or so have been
> rock solid for me, at least starting with 2.1.97 and working
> forward, although 2.1.91 and later looked pretty good.

> Looks like progress... waiting on Linus's new "looks so good it
> must be right" interrupt handler and version 2.1.103 now. :)

> --
> Robert Hyatt Computer and Information Sciences

> (205) 934-2213 115A Campbell Hall, UAB Station
> (205) 934-5473 FAX Birmingham, AL 35294-1170

I've had 2.1.103 up since early Friday and it's rock solid, 2.2 can't be very
far away.
Some good things have gone into 2.1.x, the 64 Meg "limit" has disappeared, just
plug the memory in and switch on.
One thing that caught me on the hop was "ipchains", profuse documentation,
except none of the examples fitted a simple masquerading example, when all else
had failed, there was only "forward" to work on and it did the trick. I did it
on the fly in 2.1.102, did ipchains-save, but when I rebooted on 2.1.103,
"ipchains-restore" just sat there complaining and the login prompt never
appeared, so I booted up off tomsrbt, edited my script to remove it and replaced
it with a ipchains command.
Regards
--
... Sid Boyce...Amdahl(Europe)...44-121 422 0375
-----------------------------------
Any opinions expressed above are mine and do not necessarily represent
the opinions or policies of Amdahl Corporation.

Very Computer

2.1.102 kernel

2.1.102 kernel

2.1.102 kernel

2.1.102 kernel