On 1/17/24 15:13, gene heskett wrote:
On 1/17/24 11:30, Thomas Schmitt wrote:
Hi,
after i began enumerating suspects, gene heskett wrote:
terminals scroll back memory, I purposely set this
particular terminals scrollback to 200 lines with that in mind.
How large was it set when your runs caused the OOM killer to act ?
different terminal, xfce4's is apparently unlimited but can't find it in
the config prefs.
I have a good number of xterms with 10,000 lines each. No tabs, no KDE,
but 8 fvwm "desktops" (virtual screens) full of terminal windows.
12 workspaces with 1 to 8 tabs open. 32G of main memory.
[Request to test the disks one-by-one on some other computer, whether
they bear the same serial number at all controllers in all machines.]
Not as easily tried, the other 4 are in twin mounts in another
portion of
the drive cages in this 30" tall tiger direct cage and not too readily
accessible w/o tipping the mobo out on its hinged mount.
One should raise protest at Gigastone if the disks really have the same
serial numbers. But before doing so, one would have to make sure that
it is not some weird effect of them all being plugged into that machine
at the same time.
Should not be a problem if labeled uniquely. And that's easily affected
by gparted.
One of you made the remark that seems to be the secret password.
What did finally help ? Just the shorter terminal scroll back memory ?
That, and possibly the --bwlimit=10m, giving the SSD time to keep their
stuff in one sock.
It would explain why a verbose rsync could summon the OOM killer always
around the same stage of progress. But what waste of memory would have
to happen with each of the rsync messages ?
Everything you see flying by when the -v is in the opts, and some of the
pathnames are 250-300 bytes long.
(You mentioned LABEL as a possibility. But not as actually used.)
Yes I have, repeatedly.
Its still, slowly at 10 megs a second, working.
I see in your previous mail rsync option --bwlimit=10m . But in the
same mail there is an older quote from you that --bwlimit=3m only
prolonged the time until the OOM killer appeared.
So i wonder whether it would work at a more contemporary speed.
I can't change it for testing? Boggles my mind.
A probably informative test.
But as yet not tested.
------------------------------------------------------------------------
Self-incrimination: The rest of this mail is off topic.
they gave all 7nth graders the Iowa
test in 1947, similar to the S/B IQ test but not copyrighted, there
fore a
lot cheaper, and I came out of that with an equivalent of 147.
I was tested in the 1960s but they did not tell the results to kids or
parents. We only got recommendations at which of our three types of
school
we should continue at the age of 10 or 11 years.
That I believe was the intention but one of the teachers was a
blabbermouth.
(So it was not to avoid discrimination of the dumb but rather to avoid
that pupils feel more intelligent than their teachers.)
That avoidance was untenable, in the 1st semester of my freshman year I
got thrown out of the senior physics class for correcting an erroneous
statement by the teacher that was patently at odds with Newton's 3rd law
of motion. For every action, there is an equal but opposite reaction.
Pretty basic stuff. But correcting the teacher in front of the other
students was absolutely not to be tolerated. But I felt correcting him
AND setting it straight was more important to the rest of the nominally
20 students present than any embarrassment it may have caused him.
Same with the papered EE's who can't understand that E=MV2 does not have
a speed floor, below which its doesn't work when the electron beam in a
klystron amplifier is only moving at a potential of 20,000 volts. The
problem not understood is that the amplification is obtained not from a
current variation, but a velocity variation induced by a 1 watt signal
speeding up or slowing down the passing beam as it traverses the first
cavity of 4, the next two to control the bandwidth, the last one picks
30 kilowatts back off the beam by the capacitative coupling effects as
the beam goes on thru into a copper funnel cooled by 70 gallons of very
pure water to absorb the end of that beam which takes around 125
kilowatts to generate.
I forgot to mention that 70 gallons figure is a per minute value
supplied by a 15 hp ingersol-rand pump. A semi sealed system that has a
4' wide x8' long x1.5' thick radiator supplied with external cooling air
by a another 20 horse motor. Rigged by vent louvers to control the air
flow to maintain the water above freezing. That 20 horse had the power
to blow that whole louver out into the field behind the building when
the modutrol motor that controlled that hot air exit louver failed to
open it at signon time one morning. Panic call from the remote control
site as it was only about 20F outside and the water was getting colder,
endangering the radiators freezing and $250,000 dollars worth of klystrons.
But that beams electrons have mass, another name for weight, and one
watt to slow them slows them more than 1 watt to speed them up speeds
them up, so at high power levels, the tube is effective longer in terms
of the transit time. This puts a time of flight error into the signal we
didn't know how to pre-distort for in the 1970's. A very dependable way
to generate transmitter power levels that was also not very efficient,
95% of the uhf stations that went dark in those years were bankrupted by
the power bills even at 3 cents a kw.
So there was a huge financial push to find a better method as that time
distortion would have killed hidef tv before it ever got out of the
laboritory,
And E=MV2 is as valid at 25 mph as it is at C speed, nominally 186,272
miles per second.
Yup, I understand Albert Eintein's theory. Did this help you to
understand it? I hope so.
Have a nice day :)
Thomas
Thanks for pulling the trigger of the teacher I can be, Thomas. BTW,
rsync is still toddling along.
root@coyote:~# df && free
Filesystem 1K-blocks Used Available Use% Mounted on
udev 16327704 0 16327704 0% /dev
tmpfs 3272684 1904 3270780 1% /run
/dev/sda1 863983352 22348260 797673444 3% /
tmpfs 16363420 1244 16362176 1% /dev/shm
tmpfs 5120 8 5112 1% /run/lock
/dev/sda3 47749868 680 45291180 1% /tmp
/dev/md0p1 1796382580 335104628 1369952976 20% /home
tmpfs 3272684 3756 3268928 1% /run/user/1000
/dev/sdh1 1967892164 239677008 1628178908 13% /mnt/homevol
total used free shared buff/cache
available
Mem: 32726840 3552420 307328 921164 30132056
29174420
Swap: 111902712 1536 111901176
So the copy is a little over 2/3rds done in nominally 8 hours.
Cheers, Gene Heskett.
Cheers, Gene Heskett.
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
