From mboxrd@z Thu Jan 1 00:00:00 1970 From: jens Subject: maybe stability issues Date: Wed, 29 Apr 2009 14:26:26 +0200 Message-ID: <49F84772.1060708@gis.de> Reply-To: jens@gis.de Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit To: kvm@vger.kernel.org Return-path: Received: from mail3.gis.de ([212.51.29.20]:44811 "EHLO mail3.gis.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751604AbZD2M5h (ORCPT ); Wed, 29 Apr 2009 08:57:37 -0400 Received: from localhost (unknown [127.0.0.1]) by mail3.gis.de (Postfix) with ESMTP id 6F054172797 for ; Wed, 29 Apr 2009 14:26:29 +0200 (CEST) Received: from mail3.gis.de ([127.0.0.1]) by localhost (mail3 [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 08892-01 for ; Wed, 29 Apr 2009 14:26:24 +0200 (CEST) Received: from [172.16.254.80] (unknown [95.118.178.198]) (Authenticated sender: jens) by mail3.gis.de (Postfix) with ESMTP id C863116E450 for ; Wed, 29 Apr 2009 14:26:22 +0200 (CEST) Sender: kvm-owner@vger.kernel.org List-ID: Hi, I maybe notized for me hard to reproduce stability issues with kvm and therefore I would like to ask the list for a hint how to solve the problem. The kernel I use for the smp host and guest is a 2.6.29.1 kernel.org kernel and I compiled kvm-85 with '--with-patched-kernel'. For the three guests I use the virtio block- and network-device and KVM_CLOCK. The block-devices are drbd-8.3.1 devices on top of lvm2. The smp host is equipped with 8G ram and one intel core2 quad cpu at 2.83GHz. Two of the three quests (guest1 and guest2) are assigned 2G ram and 4 vcpus, the third guest (guest3) is assigned 1G ram and 4 vcpus. To manage the guests I use virsh 0.6.2 (debian unstable). Today I noticed that guest2 was not running, I started the guest again and looked at the logfiles but I couldn't figure out why the guest had stopped. To test if a system reboot generally works I opened a console on guest3 and started 'reboot' and the guest shuts down but don't came back to life again even though 'virsh list' shows guest3 in running state. At this point I noticed that the smp host consumed much cpu power. Next I destroyed guest3 using virsh but not only guest3 was destroyed, guest1 was destroyed at the same time too. At the moment I have absolutely no idea how to figure out what went wrong because I'm new to kvm and I found no obvious reason for this behaviour. The only thing I noticed was the message "hrtimer: interrupt too slow, forcing clock min delta to 141126894 ns" on the console. Maybe I just used a unstable combination of kvm modules, virtio driver, kernel, kvm userland- and libvirt-utilities? I hope that someone on the list with more experience can point me to the right direction. For the sake of completeness I'll attach one config.xml, the three configs are almost the same. Greetings jens -- guest2 a5bf0d7c-c541-82a8-9241-fc5dbf59beb1 2096128 2096128 4 hvm /boot/vmlinuz-2.6.29.1-wickie root=/dev/vda ro console=tty0 console=ttyS0 destroy restart restart /usr/bin/kvm