Discussion:
o2 slowdown
(too old to reply)
kenbo
2008-01-24 20:48:48 UTC
Permalink
Raw Message
I have an O2 that is being used as a general image acquisition
system. It's always been a fine
machine, even with 256 megs and 16 gigs of hard drive space. It is
running Irix 6.5
A real workhorse, now getting
to be an oldtimer :-)



This week, when we went to do an image capture, using mediarecorder as
always, the machine
was slow. The mediarecorder window popped up quickly, but the icons
took several seconds
for each to come in. If you select a menu item, it takes up to 30
seconds for the menu to
drop down :-(

Nothing has changed in the hardware configuration. No unwanted tasks
are running.
Load ranges from .30 to 1.5 but nothing unusual.

I ran the diagnostics and things checked out fine. I went to the
demos
and ran several. distort crashed, atlantis crashed, pointblank ran
fine,
texcube ran slow.

What could be causing a generalized slowdown in the graphics speed?

Any thoughts would be appreciated.

kenny
Toni Grass
2008-01-25 00:17:57 UTC
Permalink
Raw Message
Post by kenbo
This week, when we went to do an image capture, using mediarecorder as
always, the machine
was slow. The mediarecorder window popped up quickly, but the icons
took several seconds
for each to come in. If you select a menu item, it takes up to 30
seconds for the menu to
drop down :-(
Did you see any entries in system log which point to SCSI or hard disk
problems? How full is your hard disk? I am sure to remember a hint in
IRIX manual to keep the used disk space in a file system less than 90%
or so.

Toni
kenbo
2008-01-25 01:13:50 UTC
Permalink
Raw Message
Post by Toni Grass
Post by kenbo
This week, when we went to do an image capture, using mediarecorder as
always, the machine
was slow. The mediarecorder window popped up quickly, but the icons
took several seconds
for each to come in. If you select a menu item, it takes up to 30
seconds for the menu to
drop down :-(
Did you see any entries in system log which point to SCSI or hard disk
problems? How full is your hard disk? I am sure to remember a hint in
IRIX manual to keep the used disk space in a file system less than 90%
or so.
Toni
Good ideas.

I checked the syslog for suspicious activity; nothing abnormal was
there.
The only unusual thing in the log was a lot of processes for bootp
being
discussed. But, this seems to be consistent over a long period of
time.
This slowdown has happened recently.

The file system on the root drive was at 95%. I cleaned up some
backup file
that were no longer needed and got it down to 45%! (Hey, I don't
clean up files
until needed :-) The effect on mediarecorder was the same.

Thanks for the ideas, I will keep looking for other things.

kenny
Toni Grass
2008-01-25 21:26:34 UTC
Permalink
Raw Message
kenbo wrote:

[....]
Post by kenbo
Thanks for the ideas, I will keep looking for other things.
Did you try 'top' to find out which program wastes CPU time?

Toni
Barutan Seijin
2008-01-27 01:22:41 UTC
Permalink
Raw Message
Post by Toni Grass
[....]
Post by kenbo
Thanks for the ideas, I will keep looking for other things.
Did you try 'top' to find out which program wastes CPU time?
I second that. You may very well have a run away process. I had a
similar mysterious slowdown on an OSX machine. Checking top and a deft
use of kill did the trick. Give it a try.
--
***@gmail.com
kenbo
2008-01-27 15:25:58 UTC
Permalink
Raw Message
Post by Barutan Seijin
Post by Toni Grass
[....]
Post by kenbo
Thanks for the ideas, I will keep looking for other things.
Did you try 'top' to find out which program wastes CPU time?
I second that. You may very well have a run away process. I had a
similar mysterious slowdown on an OSX machine. Checking top and a deft
use of kill did the trick. Give it a try.
--
top revealed nothing other than top and mediarecorder at the top of
the list :-(

I rebooted the machine (a painful mental decision since it had been up
for
over a year :-), but the problem was still there.

I am particularly puzzled by the various demos that now crash, even
though they used to run fine. It's an interesting problem but I am
stuck on what to do next. Well, not exactly---if I cannot solve it,
I
can probably replace it with a similar set of hardware. But,
that's avoiding the problem not solving it :-)

kenny
Toni Grass
2008-01-27 15:42:20 UTC
Permalink
Raw Message
kenbo wrote:

[...]
Post by kenbo
I am particularly puzzled by the various demos that now crash, even
though they used to run fine. [...]
wild guess: bad sectors/ tracks on hard disk. Did you try the hardware
test from boot prom?

Toni
Gary Heston
2008-01-27 15:51:54 UTC
Permalink
Raw Message
In article <11a98c44-fd93-4d8b-978a-***@f47g2000hsd.googlegroups.com>,
kenbo <***@gmail.com> wrote:
[ ... ]
Post by kenbo
top revealed nothing other than top and mediarecorder at the top of
the list :-(
I rebooted the machine (a painful mental decision since it had been up
for
over a year :-), but the problem was still there.
I am particularly puzzled by the various demos that now crash, even
though they used to run fine. It's an interesting problem but I am
stuck on what to do next. Well, not exactly---if I cannot solve it,
I
can probably replace it with a similar set of hardware. But,
that's avoiding the problem not solving it :-)
Before buying replacement hardware, try your drive in another O2; if
it's still slow, it's not a hardware problem.

Have you checked the system logs for error messages?


Gary
--
Gary Heston ***@hiwaay.net http://www.thebreastcancersite.com/

We live in amazing times, when one person can invent both the Internet
and global warming, then get awarded a "peace prize".
Walther Mathieu
2008-01-30 13:10:42 UTC
Permalink
Raw Message
Post by kenbo
top revealed nothing other than top and mediarecorder at the top of
the list :-(
I rebooted the machine (a painful mental decision since it had been up
for
over a year :-), but the problem was still there.
I am particularly puzzled by the various demos that now crash, even
though they used to run fine. It's an interesting problem but I am
stuck on what to do next. Well, not exactly---if I cannot solve it,
I
can probably replace it with a similar set of hardware. But,
that's avoiding the problem not solving it :-)
kenny
First, I´d recommend a full system backup, then a disk clone.
SCSI disks tend to die slowly, not all of a sudden.

Is there any remote directory mounted on your O2 eventually?
An overcrowded NFS directory may slow down the entire desktop.

Demos that suddenly crash suggest a more serious problem...
is some RAM aging/performing a lot of ECC (just wildly guessing)?

Personally, I´d suspect bootp messages - do you boot from a server?
Do other IRIX machines boot from your O2?
Maybe your system was compromised somehow - try a google search
Post by kenbo
By sending a specially crafted packet to a BootP daemon a malicious
user can cause the exeution of arbitrary code (or if it failes, the
BootP daemon will crash and stop responding).
Linux, >OpenBSD, FreeBSD, Solaris, Irix, Digital Unix, and many others,
and it >been confirmed that on OpenBSD is vulnerable to the remote
execution, >while FreeBSD and Linux (RedHat, Slackware and Debian) are
not >vulnerable.

Try to run the machine standalone to see if it behaves different.

walther
josehill
2008-01-31 06:17:52 UTC
Permalink
Raw Message
Before messing with hardware, I'd try to launch one of the broken
demos via the command, running it through the "process activity
reporter," also known as the "par" command. It will log every call
made by the program, often revealing if there is a problem opening a
file (like perhaps a damaged shared library or config file) or a
directory. The output can be long and a little daunting, but look for
lines with the words "open" or "fail" in them for possible clues.

I haven't run it in a while, but IIRC, the syntax for a par cpmmand
line with reasonable options is
par -i -s -SS -o par_output.txt <insert your command-line here>

For example, to learn how "ls -al" behaves, you might use
par -i -s -SS -o par_output.txt ls -al

Read the man page for more details.
Post by kenbo
I am particularly puzzled by the various demos that now crash, even
though they used to run fine. It's an interesting problem but I am
stuck on what to do next. Well, not exactly---if I cannot solve it,
I
can probably replace it with a similar set of hardware. But,
that's avoiding the problem not solving it :-)
kenny
Loading...