From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:49656) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1S4bnn-0005eI-Rk for qemu-devel@nongnu.org; Mon, 05 Mar 2012 12:34:29 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1S4bnh-00022G-Bh for qemu-devel@nongnu.org; Mon, 05 Mar 2012 12:34:27 -0500 Received: from mail-pz0-f46.google.com ([209.85.210.46]:63888) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1S4bnh-00021k-3D for qemu-devel@nongnu.org; Mon, 05 Mar 2012 12:34:21 -0500 Received: by dajr28 with SMTP id r28so6691453daj.33 for ; Mon, 05 Mar 2012 09:34:18 -0800 (PST) Sender: Paolo Bonzini From: Paolo Bonzini Date: Mon, 5 Mar 2012 18:33:54 +0100 Message-Id: <1330968842-24635-1-git-send-email-pbonzini@redhat.com> Subject: [Qemu-devel] [PATCH v3 0/8] Mirrored block writes List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: kwolf@redhat.com, fsimonce@redhat.com, eblake@redhat.com, stefanha@linux.vnet.ibm.com, lcapitulino@redhat.com v3 comes with a new QMP command drive-mirror, an improved blockdev-snapshot-sync that just reuses transaction functionality, and a new image creation mode enum. I also tested that the command can be used to migrate without shared storage. Tested with the following scenarios: a) mirror only 1) create base.qcow2 and starat QEMU with it 2) Execute the following QMP command { "execute": "qmp_capabilities" } { "execute": "transaction", "arguments": {'actions': [ { 'type': 'drive-mirror', 'data' : { 'device': 'ide0-hd0', 'target': '/home/pbonzini/mirror.qcow2' } } ] } } { "execute": "cont" } 3) hibernate the guest (this requires an IDE disk and -cpu kvm64,-kvmclock) 4) restart the guest with mirror.qcow2 b) Same as (a) with drive-mirror command. c) streaming to new image 1) start QEMU with an existing image test.img 2) execute the following HMP commands drive_mirror -s ide0-hd0 /home/pbonzini/mirror.qed qed block_stream ide0-hd0 3) shut down the guest (sorry, this took a looong time and I forgot to hibernate here) 4) move away the base image 5) restart the guest with mirror.qed d) atomic snapshot+mirror (QMP only): 1) start QEMU with an existing image test.img 2) Execute the following QMP command { "execute": "qmp_capabilities" } { "execute": "transaction", "arguments": {'actions': [ { 'type': 'blockdev-snapshot-sync', 'data' : { 'device': 'ide0-hd0', 'snapshot-file': '/home/pbonzini/base.qcow2' } }, { 'type': 'drive-mirror', 'data' : { 'device': 'ide0-hd0', 'target': '/home/pbonzini/mirror.qcow2' } } ] } } { "execute": "cont" } 3) hibernate the guest (this requires an IDE disk and -cpu kvm64,-kvmclock) 4) check that mirror.qcow2 has test.img as the base 5) restart the guest with base.qcow2 6) restart the guest with mirror.qcow2 v2->v3: replace reuse argument with NewImageMode enum and mode argument; renamed all commands and actions; implemented blockdev-snapshot-sync in terms of group snasphots; added new drive-mirror command and HMP equivalent. Marcelo Tosatti (1): Add blkmirror block driver Paolo Bonzini (7): fix format name for backing file qapi: complete implementation of unions rename blockdev-group-snapshot-sync add mode field to blockdev-snapshot-sync transaction item qmp: convert blockdev-snapshot-sync to a wrapper around transactions add mirroring to transaction add drive-mirror command and HMP equivalent Makefile.objs | 2 +- block/blkmirror.c | 239 +++++++++++++++++++++++++++++++++++++++++++++ blockdev.c | 232 ++++++++++++++++++++++++------------------- hmp-commands.hx | 31 +++++- hmp.c | 34 ++++++- hmp.h | 1 + qapi-schema-test.json | 10 ++ qapi-schema.json | 119 ++++++++++++++++++---- qmp-commands.hx | 105 +++++++++++++++----- scripts/qapi-types.py | 5 + scripts/qapi-visit.py | 31 ++++++- test-qmp-input-visitor.c | 18 ++++ test-qmp-output-visitor.c | 34 +++++++ 13 files changed, 707 insertions(+), 154 deletions(-) create mode 100644 block/blkmirror.c -- 1.7.7.6