headless and I control via ssh
over local lan.
I can leave the current debian setup on the disk or wipe it.
Heitzso
heitzso at growthmodels dot com
top has been failing for a week or so on my ss20 box running
2.4.22 compiled single cpu. Error message is
"top: failed /proc/stat read"
strace top yields (last few lines)
open("/proc/stat" ... = 5
fstat64(5, ...) = 0
mmmap(
_llseek(
read(5, ...) = 243
write(1, "..." = 317
ioctl(0,
he moment about
this problem (the archives search server is not
responding.)
Heitzso
email to the list)
- umounts of my NFS backup volume is failing, even
after the hung tar is killed
BTW, al of these pieces were working a couple of weeks
ago before the upgrade that forced me to bump from 2.2.25
to 2.4.22.
Suggestions?
Thanks,
Heitzso
were breaking instead of the kernel?
Should I sit tight for some patchset to come down the
pike in a future 2.4 or 2.6 kernel?
Also, and this is very strange, 'top' stopped working
in 2.4.22, though worked great in 2.2.25. I'd appreciate
any clues re getting it to work again.
Thanks,
Heitzso
r
box and see if the network is up, though that assumes all eth
settings okay with the other port.
And this is just a wild guess. If I'm right, setting alias between
the eth0 and your hme card will fix things.
Heitzso
Steve King wrote:
I have experienced a weird problem with the networ
glitches elsewhere have been fixed.
Is that true?
Also, if I build a custom 2.4.? SMP kernel what compiler should I be using?
And should the source be vanilla or debian patched?
Thanks,
Heitzso
ystem. I can't spend days thrashing this again.
If I have to, I can go to single cpu kernel (they work on my system) but
really want to get the SMP rolling.
What is the current recommended most-likely-to-work kernel/patch/gcc
setup to run SMP on dual hypersparc 150 ss20?
Thanks,
Heitzso
W
and locales. Comments on safest path appreciated. (Wait and
retry tomorrow? Use -f?)
Thanks,
Heitzso
s running 2.4.22-ac3, both are
on 'unstable' so there's the steady thrash of packages,
and I've been trying to push key packages over to
-mv8 packages on the sparc box to get the integer math speedup
for some python/openssl/whatever packages.
Thanks
Heitzso
, yes, the decreased time is why I've been fighting
it -- I want that 2-3 second ssl setup time. lsh gave
it to me w/o compiling from source for ssh2 and cvs-over-ssh2
but email-over-tls-libssl is still the agonizingly slow 10 seconds.
Michael Andersen wrote:
Heitzso wrote:
Problem
up ssh via openssl. My email TSL still routes via openssl/libssl
and takes the 10 seconds, but otherwise ...
Heitzso
Andrew Sharp wrote:
On Thu, Oct 16, 2003 at 12:16:24PM -0400, Heitzso wrote:
Eric Wong wrote:
I used very conservative optimizations, but nevertheless, it looks go
some speedup over -mv7 flavor distributed.
I'm still trying to get a ssh logon to my ss20 to drop
down to a few seconds. Somewhat frustrating for every
cvs/email/ssh (all using ssl) to take about 10 seconds for
the ssl to setup.
Heitzso
I'm still fighting a debian libssl0.9.7 build trying to execute
a 64 bit program on my 32 bit sparcstation 20 system
(current as of this morning debian source).
I'm watching other emails talk about 32/64 confusions
in sparc ports/builds.
I thought I'd 'apt-get remove' useless 64 bit packages
off
in ../ If
that fails also, contact the upstream author.
As I understand it, apt-build is pretty much just a wrapper for those
steps.
I'm glad you're having at least some success, though! I've not had
the time to piddle with it this month like I wanted.
/dev/idal
Heitzso
rogram is still complaining
so I'll reboot and cross my fingers. Any other
ideas? Anyone else have this problem?
Thanks,
Heitzso
I haven't studied the guts of what's going on. I'm just reporting
what happens with apt-build on 32 bit sparc (mv8) system.
Please let me know if there's some test or system check
I should run to help clarify this problem.
BTW, otherwise apt-build tends to work (I've used to
build/install maybe
ompile it's harder to hold the pieces
together.
Does anyone have a suggestion re getting this package
to compile with apt-build?
Thanks,
Heitzso
? I only had CD 1 non-us debian sparc
stable on hand and the primitive shell that's provided
in that environment was frustrating. Are there
knoppix/gnoppix like rescue CDs around for sparc?
Thanks,
Heitzso
ries somewhere along the
way that breaks everything. We'll see.
Heitzso
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hello all
I am interested at recompiling all packages that i have installed on my ultra5
currently running unstable. (yeah its a slow box, but its idle too)
I am at the
parc
Is this catching all possible hypersparc optimizations?
Thansk!
Heitzso
so, given that it was about midnight, I thrashed it
back to the original state without noting the exact
wording of the eternal error loop. I apologize for
not grabbing the exact text.
Heitzso
Eric Brower wrote:
Heitzso wrote:
I have an sbus antares (actually a qlogic)
fast/wide scsi
few days before
I get to it.
My main thrash is trying to figure out why
postfix-tls-openssl-saslauthd-pam stopped sending out the sasl
auth modes recently. I'm rigging the ss20 to be a hopefully
stable SOHO edge box with mail server.
Heitzso
Eric Brower wrote:
Yep-- missing the "int
ing controller etc. message.
Thanks for digging up the message in the driver.
Still no scsi controller. I'll slide on it since
I'm not in a position to thrash the code.
Heitzso
Eric Brower wrote:
Heitzso wrote:
I have an sbus antares (actually a qlogic)
fast/wide scsi2 controller in
2.2.X
kernels. I've scanned kernel doc and googled
around but haven't found the right clue.
Any ideas would be appreciated. It would be real
nice to get the speed bump from this controller!
Thanks,
Heitzso
because the dist-upgrade will rip
amavisd-new out of your system. Between that and having
sasluathd ripped out by a prior 'apt-get dist-upgrade'
my mail server is feeling rather vulnerable.
Just some FYIs while I wonder about my sanity.
Heitzso
l.org. Since
this 2.2.25 kernel doesn't crash at INIT on the way up
I'ld guess it isn't the INIT code per se. ??
Just trying to add data points.
Thanks,
Heitzso
I have a sparc 20, 2x 150 MHz HperSparcs, and 2.4.21-SMP, vanilla, no patch,
ext3 all over.
My machine always boots up t
t;unimplemented sparc system call" error messages
popped up on the console. The unimplemented sparc
calls were 32, 53, 69, and 89. Should I be
worried? (said nervously ...)
Heitzso
limitation and how to
measure what's generated by the compile
I'ld appreciate the pointer.
Thanks,
Heitzso
nicom watching the ttya serial port)
bootlogd complained of not being able to access
/dev/tty0
Any suggestions would be greatly appreciated,
Heitzso
ox.
Any pointers would be appreciated. I've googled
on most word combos but am not seeing recent
references to the hypersparc/smp combo.
Thanks!
Heitzso
on of the kernel. A simple 'make' builds
vmlinux and I (believe, haven't tried) can work
from there. NOTE I assume this is not related
to sparc architecture but have not tracked it down.
Finally, I haven't tried running with this new
2.4.23-pre5 kernel yet. That's next.
Heitzso
32 matches
Mail list logo
