Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Jun 2018 18:09:11 -0700
From:      bob prohaska <fbsd@www.zefox.net>
To:        Mark Millard <marklmi@yahoo.com>
Cc:        Trev <freebsd-arm@sentry.org>, freebsd-arm <freebsd-arm@freebsd.org>, bob prohaska <fbsd@www.zefox.net>
Subject:   Re: GPT vs MBR for swap devices
Message-ID:  <20180622010911.GA98112@www.zefox.net>
In-Reply-To: <1D86911D-20D1-494A-822B-1C07C5598CB1@yahoo.com>
References:  <25F1A4BA-FBFC-4C32-85DD-5F5BA71A2B1A@yahoo.com> <20180620023253.GA89924@www.zefox.net> <a232ed45-a9a9-1017-72ed-720a6c7a8f03@sentry.org> <1D86911D-20D1-494A-822B-1C07C5598CB1@yahoo.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jun 21, 2018 at 07:16:19AM -0700, Mark Millard wrote:
> 
> 
> 
> It looks like the latest from Bob P. is that when his /dev/da0
> is not used for swap (and so has far less I/O) his builds finish,
> apparently without logging errors:
> 
> 
A little more testing has been completed. 

With enough swap not on da0 -j4 buildworld finishes. It does not seem to 
matter what the swap is on: SD flash, USB flash or USB mechanical. With 
any swap on da0 -j4 buildworld does not finish because of intervention
by OOMA (the out of memory assassin). Occasionally, OOMA runs amok,
killing processes not obviously consuming lots of memory. 

With not enough swap not on da0 the result is somewhat surprising. 

In three trials of using 1 GB of swap on the SD card, the machine reported it was out
of swap, the OOM assassin did not come to the rescue and eventually the machine
stalled. The log files of gstat and swapinfo data are at
http://www.zefox.net/~fbsd/rpi3/swaptests/newtests/1gbsdflash_swapinfo/1gbsdflash_swapinfo.log
http://www.zefox.net/~fbsd/rpi3/swaptests/newtests/1gbsdflash_swapinfo/newtest/2nd1gbsdflash_swapinfo.log
and
http://www.zefox.net/~fbsd/rpi3/swaptests/newtests/1gbsdflash_swapinfo/newtest/3rdtest/3rd1gbsdflash_swapinfo.log

Each directory also holds what I could gather from the console and buildworld.

The errors on the console seem to suggest a loss of communication with da0. Could 
something like usbdump be used to shed more light on what's going on? 

Thanks to all for reading!

bob prohaska
 



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20180622010911.GA98112>