|
From: | Peter Lieven |
Subject: | Re: [Qemu-block] qemu exit with pending I/O |
Date: | Mon, 27 Apr 2015 14:23:25 +0200 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0 |
Am 27.04.2015 um 14:11 schrieb Paolo Bonzini:
On 27/04/2015 13:51, Peter Lieven wrote:The reason to do this, is that while I/O is pending you do not know if the I/O has been completed or not.But if I consider a raw image, quitting without a clean shutdown can also leave an inconsistent state?But you know that all I/O submitted to disk has been handled. The bdrv_drain_all() avoids that some leftover I/O mysteriously pops up while another VM, or another instance of this VM, is using the image. That's much worse.
Can this happen with a raw image used by a single VM? Would it be an option to have a quit command with lets call it "noflush" flag? I consider the case that my storage has died. In the iSCSI case all vServers will end up in the now async reconnect loop. In such a case I might need to terminate the vServers and want to restart them later from a backup. It would be interesting to be able to terminate the vServer with a command via hmp/qmp wihtout the need to send a SIGKILL from external. Peter
[Prev in Thread] | Current Thread | [Next in Thread] |