From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marcelo Tosatti Subject: Re: Hangs Date: Fri, 21 Nov 2008 20:32:37 +0100 Message-ID: <20081121193237.GA22203@dmt.cnet> References: <200811192243.mAJMhgFU005350@versecorp.net> <20081120171057.GA13638@versecorp.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: kvm@vger.kernel.org, rl@brabbel.net To: chris@versecorp.net Return-path: Received: from mx2.redhat.com ([66.187.237.31]:41888 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752840AbYKUWez (ORCPT ); Fri, 21 Nov 2008 17:34:55 -0500 Content-Disposition: inline In-Reply-To: <20081120171057.GA13638@versecorp.net> Sender: kvm-owner@vger.kernel.org List-ID: On Thu, Nov 20, 2008 at 09:10:57AM -0800, chris@versecorp.net wrote: > On Wed, Nov 19, 2008 at 02:43:42PM -0800, chris@versecorp.net wrote: > > Thanks for the responses, > > > > I'm not sure if my problem is the same as Roland's, but it definitely sounds > > plausible. I had been running ntpdate in the host to synchronize time every hour (in a cron job), so it sounds as if we could be seeing the same issue. > > > > Actually, with ntpdate taken out of crontab, I'm still seeing periodic > hangs, so it's either a different problem or I'm hitting it in a > different manner. > > OK, I installed kvm-79 and kernel 2.6.27.6, and here's the the kvm-stat output > with 1 guest hung and 3 more operational: > If I shut down the 3 operational guests leaving just the hung guest, the > kvm-stat output is all 0s: > > efer_relo exits fpu_reloa halt_exit halt_wake host_stat hypercall insn_emul insn_emul invlpg io_exits irq_exits irq_windo largepage mmio_exit mmu_cache mmu_flood mmu_pde_z mmu_pte_u mmu_pte_w mmu_recyc mmu_shado nmi_windo pf_fixed pf_guest remote_tl request_i signal_ex tlb_flush > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 So the guest is not actually running here, which means its QEMU that its hanging at. > The hung guest in this case was run with this command: > > sudo /usr/local/bin/qemu-system-x86_64 \ > -daemonize \ > -no-kvm-irqchip \ > -hda Imgs/ndev_root.img \ > -m 1024 \ > -cdrom ISOs/ubuntu-8.10-server-amd64.iso \ > -vnc :4 \ > -net nic,macaddr=DE:AD:BE:EF:04:04,model=e1000 \ > -net tap,ifname=tap4,script=/home/chris/kvm/qemu-ifup.sh \ > >>& Logs/ndev_run.log > > > I should also mention that when the guest is hung, I can still switch > to the monitor with ctrl-alt 2. So, at least it's a little bit alive. In coma perhaps. > I've also noticed that the behavior with the hung guest is slightly > different on kvm-79 than it was earlier. When the guest hangs, the kvm > process in the host doesn't spin at 100% busy any longer - the guest is > just unresponsive at both the network and VNC console. > Also, I've noticed that if I reset the guest from the monitor, the > guest will boot up again, and I can get through to it on the network, > but strangely, the mouse and keyboard will still be hung at the > VNC console (except that I can still switch back and forth to the > monitor). > > Hope some of this helps, let me know if you need to me to provide any > other troubleshooting info. $ gdb -p pid-of-qemu (gdb) info threads Print the backtrace for every thread with: (gdb) thread N (gdb) bt