From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:43548) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1eHSlC-0007FY-HG for qemu-devel@nongnu.org; Wed, 22 Nov 2017 06:00:11 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1eHSl8-0003e0-S2 for qemu-devel@nongnu.org; Wed, 22 Nov 2017 06:00:06 -0500 Received: from mx1.redhat.com ([209.132.183.28]:18170) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1eHSl8-0003dg-ME for qemu-devel@nongnu.org; Wed, 22 Nov 2017 06:00:02 -0500 Date: Wed, 22 Nov 2017 10:59:58 +0000 From: "Dr. David Alan Gilbert" Message-ID: <20171122105957.GC2443@work-vm> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Subject: [Qemu-devel] postcopy test under load List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: peter.maydell@linaro.org Cc: qemu-devel@nongnu.org Hi Peter, I was trying to reproduce the occasional travis postcopy test case failures you were seeing; I can reproduce them but only with tcg - which makes me wonder if the problem is in the test or actually in tcg. My test is: export QTEST_QEMU_BINARY=wherever/try/x86_64-softmmu/qemu-system-x86_64 cd whereever/try for NUM in $(seq 1 100) do echo 'while true; do ./tests/migration-test || break; done' | taskset 2 bash & done so all of the instances are pinned to one core using taskset. I had 100 instances of kvm running overnight (x86-64) and didn't hit any problems. 30 instances of tcg hit the problem in 5 minutes (x86-64 on x86-64). Dave -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK