Date: Tue, 21 Nov 2017 13:36:44 -0800 From: Mark Millard <markmi@dsl-only.net> To: bob prohaska <fbsd@www.zefox.net>, Freebsd-arm <freebsd-arm@freebsd.org> Subject: Re: ssh sessions close spontaneously on rpi2 Message-ID: <B7232214-163C-4E60-9729-6A42B02FA584@dsl-only.net> In-Reply-To: <FB9B37A4-CD5B-45F7-85A5-9AF1E6174F86@dsl-only.net> References: <20171118165638.GA47956@www.zefox.net> <20171120165832.GA55836@www.zefox.net> <FB9B37A4-CD5B-45F7-85A5-9AF1E6174F86@dsl-only.net>
index | next in thread | previous in thread | raw e-mail
On 2017-Nov-21, at 1:15 PM, Mark Millard <markmi at dsl-only.net> wrote: > On 2017-Nov-20, at 8:58 AM, bob prohaska <fbsd at www.zefox.net> wrote: > >> Can anybody reproduce the following behavior on RPI2 ? >> >> Boot the machine multi-user and start two ssh login sessions. >> >> On one, start a top session, as a placeholder. On the second, >> su to root, cd to /usr/ports and run >> make -j8 clean > clean.log & >> >> Next, try to open another (this makes three) ssh connection. Log in, >> and just watch. >> >> On my test box at FreeBSD 12.0-CURRENT (RPI2) #10 r325987 the connection >> fails within a couple of minutes; the login session terminates without >> explanation and the connection closes. >> >> No errors are on the serial console, /var/log/messages records nothing >> related. Ssh sessions started prior to starting make clean seem unaffected. > > Given the messages that I've been getting but you have not, > I figure my attempting this under my current conditions > would not be effective. > > But I've another oddity that you might try: > > A) I booted and ran top on the serial console ( top -CaePosize ) > B) I logged in 4 ssh sessions and had each do "openssl speed" > > top gets an unexpected result: CPU 2 shows as 100% idle and the > others 0.0% idle. > > The process CPU column agrees, in that that the total is > around 300% instead of around 400%. A couple of the > tend to show closer to 50% as long as 4 are running. > At 3 running it goes back to each being near 100%. > > # uname -apKU > FreeBSD rpi2 12.0-CURRENT FreeBSD 12.0-CURRENT r325700M arm armv7 1200053 1200053 > > (I normally run a non-debug kernel that has debug symbols.) > > I may try jumping to -r325997 . > I rebooted and ended up stuck at Releasing APs. I cut and restored power and booted. And now all the cores seem to be in use. Still -r325700 . Not easily repeatable from what I can tell. A question for your test: what if you substitute computation-bound processes for your -j8 make? Do you get different results? === Mark Millard markmi at dsl-only.nethelp
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?B7232214-163C-4E60-9729-6A42B02FA584>