RPI3 swap experiments ["was killed: out of swap space" with: "v_free_count: 5439, v_inactive_count: 1"]

[bob prohaska]

On Wed, Aug 08, 2018 at 08:37:35PM -0700, John Kennedy wrote:
> On Wed, Aug 08, 2018 at 02:02:36PM -0700, Mark Millard via freebsd-arm wrote:
> > ... My understanding is that some other folks involved in the message
> > exchange replicated the OOM kills without having evidence of issues
> > like Bob P. has with large latencies for his hardware. The latencies
> > may be sufficient but not necessary.
> > 
> > Getting some of the other folks to run your tests would likely be
> > good. ...
> > Sounds like anyone else running tests should also have all 3
> > changes in place.
> > 
> > I've CC'd some folks from the exchange that I think were running
> > some tests.
> 
> My system is almost done crunching.  Applying patches and tweaks is fine, but
> I think Bob P has some other reporting tweaks going that we'd want to replicate
> so we have more than a fail/no fail.
> 
I've gathered up what I've been given and put it in
http://www.zefox.net/~fbsd/rpi3/swaptests/tools/
in hopes it'll be useful.

One thing I haven't figured out is how to organize the test result tree
in a simple way. At this point it's /revision/swapconfig/patch1/patch2/.../...
but that gets very messy fast, and ambiguous if folks don't standarize 
terminology (swap configs in particular). 

> I tend to be in the no-fail, which is more data I guess.  I also have a much
> older (slower) SD card that I could put back into service and see if that
> tweaks me a more towards the maybe-fail camp.

It seems a worthwhile experiment, but it'll complicate the test tree even more 8-)
This _can't_ be a new problem, how do folks deal with it?

Thanks for reading,

bob prohaska



> _______________________________________________
> freebsd-arm at freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-arm
> To unsubscribe, send any mail to "freebsd-arm-unsubscribe at freebsd.org"