[Discuss] I need a recommendation for a Linux-compatiblehigh-endPCbuilder

[Cy Schubert]

Hmm. What you wrote in your last email was so off-base for any VM subsystem I doubt any kernel developer would have told you that. I'm sure you misunderstood.

All VM implementations maintain a free pool which when the pool reaches a given threshold wakes the pageout thread. This is an oversimplification. If you had lots free and high paging rates (which you didn't say) then it's either a tuning issue or bug.

However more likely, your computer was memory constrained, pages were paged out. The memory shortage was relieved and your computer had lots of free memory and lots of swap used. So big deal. Pages aren't paged in unless they are needed in memory again. The whole process is called demand paging. It's normal and it's why they didn't solve anything. From the sound of it there was no problem.

Anyhow this discussion is accomplishing nothing. Let's just call it a day.

---
Sent using a tiny phone keyboard.
Apologies for any typos and autocorrect.
Also, this old phone only supports top post. Apologies.

Cy Schubert
<Cy.Schubert at cschubert.com> or <cy at freebsd.org>
The need of the many outweighs the greed of the few.
---

-----Original Message-----
From: Alan W. Irwin
Sent: 29/01/2018 22:00
To: Cy Schubert
Cc: discuss at vlug.org
Subject: RE: [Discuss] I need a recommendation for a Linux-compatiblehigh-endPCbuilder

On 2018-01-29 19:27-0800 Cy Schubert wrote:

> Really? Where did you learn this [my knowlege concerning special
memory for Linux]?

It is who (as in "who you know") rather than where.  Just to name drop
a bit, in this case, my collaborator for the timephem project is David
Howells who happens also to be a Linux kernel developer employed by
RedHat.  So when my Linux system developed those symptoms (acting like
there was a memory leak with everything being executed from SWAP, but
top and free showed plenty of free memory left in RAM), I quickly went
to David with this peculiar problem even though it had nothing to do
with our joint project.  He and his kernel colleagues had never heard
of such a problem with the Linux kernel, but he immediately zeroed in
on the possibility of a Linux special memory leak, told me what to
look up in /sys/kernel to attempt to verify the issue, and agreed what
I found there was absolutely wrong with one kind of special memory
being much too large.  From then on, we monitored the amount of that
particular kind of special memory starting from boot and quickly
proved that there was a slow but relentless leak occurring for it.
Unfortunately, (since I got the impression there was some high-power
Linux kernel help ready to tackle the issue) nothing more could be
done to debug the problem since nobody could find a way to reproduce
the symptoms my system was showing.  But that was an old kernel and
moving to the next kernel fixed the issue for me so we all let it
slide.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://vlug.org/pipermail/discuss_vlug.org/attachments/20180129/45a14c24/attachment.html>