Hello everyone:
Yesterday afternoon was quite interesting as piece by piece I turned off
the executing programs, but the random sleep pattern remained. I even
stripped all of akonadi from my system, as I didn't want to wade through
a mess.
FInally I started killing all my magma programs .. 1 by 1. Nothing changed.
I did notice that as I terminated running programs the disk access light
started flashing faster and faster (repetitively)
I finally killed a reliable C program that I had running in the
background for many years (as I am able on a live system) and when I
killed that program, the disk access light suddenly started blinking
like a warning light, very repetitive, like a flasher.
At this point I began to wonder if perhaps one of my disk drives were
failing, so I patched the system to latest levels and then backed up
everything all the while hoping that if a disk crash was impending, I've
be able to save most everything. The updates and backups went fine.
Well, the disks are okay, I went to the smart monitor section and the 3
drives are healthy.
I rebooted twice and the problem seems to have disappeared.
I still am wondering what brought all this on?
Randall
On 12/2/23 11:40, Ted Mittelstaedt wrote:
I suspect a clue to your issue is in your post, - see the following:
https://stackoverflow.com/questions/639540/how-much-memory-can-a-32-bit-process-access-on-a-64-bit-operating-system
I suspect that somewhere in that mess you might have a program that's not 64
bit clean.
Just for grins try defining an 8GB ramdrive as your swapfile and I'll bet the
freezing disappears.
Ted
-----Original Message-----
From: PLUG <[email protected]> On Behalf Of American Citizen
Sent: Saturday, December 2, 2023 10:53 AM
To: Portland Linux/Unix Group <[email protected]>
Subject: [PLUG] strange system behavior - inotify problem?
Hello:
I have a very interesting problem occurring my opensuse Leap 15.5 linux system.
I am crunching number theory modules, using magma to do descents on elliptic
curves. I do open a new konsole screen and do a tail on the magma results, in
real time. There are 6 magma programs executing in this way at the same time.
I also have 3 monitoring programs, but they are set up as a bash script file
with a while [ 1 ] loop, but each has a sleep nnn seconds at the end of the
loop.
I was getting occasionally the tail warning message, no files left for tail,
using polling instead.
I installed the inotify-comment bash script from the stackexchange web page
https://unix.stackexchange.com/questions/15509/whos-consuming-my-inotify-resources.
Running this script does not show anything unusual, I am only using 402 watches
in 131 instances. I did up the values in the sysctl.conf file to large sizes,
so I don't suspect that the problem is caused by this.
The system strangely halts the start of a new process, for up to 20 secs at a
time, for example, I tried to open up a new Konsole screen and the window opens
up, but then sits there for a period of time before completing the interior of
the screen.
Another example, while typing this email, I made a mistake and had to back up,
the moment I hit the backspace key, everything froze for about
10 secs or so, before resuming.
Third example, one of my monitoring screens suddenly went into uninterruptable
mode, trying using control-C did not work, I had to terminate the Konsole
screen and restart the bash script over again in a new window.
Right now I don't know quite what to do? This is slowing things way down for
me, while the system seems fine, I am only using 4 gigs of the 32 gigs system
RAM, so that should not be a problem.
Puzzled -
Randall
