From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:48968) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ULMVO-0004WH-0w for qemu-devel@nongnu.org; Thu, 28 Mar 2013 19:45:16 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ULMVM-0003V6-U4 for qemu-devel@nongnu.org; Thu, 28 Mar 2013 19:45:13 -0400 Received: from mail-ee0-f43.google.com ([74.125.83.43]:63651) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ULMVM-0003V2-No for qemu-devel@nongnu.org; Thu, 28 Mar 2013 19:45:12 -0400 Received: by mail-ee0-f43.google.com with SMTP id e50so25676eek.30 for ; Thu, 28 Mar 2013 16:45:12 -0700 (PDT) Sender: Paolo Bonzini Message-ID: <5154D604.4000600@redhat.com> Date: Fri, 29 Mar 2013 00:45:08 +0100 From: Paolo Bonzini MIME-Version: 1.0 References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH] Make guest OS bootable when hardware failure happens in log disk List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Seiji Aguchi Cc: "kwolf@redhat.com" , "peter.maydell@linaro.org" , "aliguori@us.ibm.com" , "mst@redhat.com" , "qemu-devel@nongnu.org" , "dle-develop@lists.sourceforge.net" , "hdegoede@redhat.com" , "kraxel@redhat.com" , Tomoki Sekiyama , "afaerber@suse.de" Il 28/03/2013 20:43, Seiji Aguchi ha scritto: > [Problem] > Currently, guest OS's messages can be logged to a local disk of host OS by creating chadevs with options below. > -chardev file,id=charserial0,path= -device isa-serial,chardev=chardevserial0,id=serial0 > > When a hardware failure happens in the disk, qemu-kvm can't create the chardevs. > In this case, guest OS doesn't boot up. > > Actually, there are users who don't desire that guest OS goes down due to a hardware failure of a log disk only. > Therefore, qemu should offer some way to boot guest OS up even if the log disk is broken. > > [Solution] > This patch skips error checks in case where opening a log file and creating chardev fail. This kind of policy does not belong in QEMU. QEMU should always fail if it sees a problem. Your management tool should just open the file outside QEMU, and then use either /dev/fd/NN or /dev/null as the destination. Paolo