RPI3 swap experiments

[bob prohaska]

OOMA is still killing processes in -j4 buildworld sessions for no obvious reason
when using mixed USB/microSD swap.  The most recent experiment is with r336877 
rebuilding itself from a clean start.

The various log files are at 
http://www.zefox.net/~fbsd/rpi3/swaptests/r336877/
The OOMA kills occur around two hours after
the worst read/write delays, which are in the low tens
of seconds. Perhaps most curiously, the long delays
don't appear to involve swap partitions. 

Similar problems now seem present with the RPI2 on
11-stable. The first failure was with r335398 trying 
to compile 336871. Buildworld has been backed down to 
-j2 and restarted in the hope it'll eventually succeed.
In this particular case all swap is on USB, in a single
2 GB partition.

It would be most interesting to see what happens if OOMA
could be turned off. Is that possible? 

Thanks for reading,

bob prohaska