From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:59386) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZWa3u-0004KD-7e for qemu-devel@nongnu.org; Mon, 31 Aug 2015 21:08:35 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ZWa3r-0007W2-0x for qemu-devel@nongnu.org; Mon, 31 Aug 2015 21:08:34 -0400 Received: from szxga03-in.huawei.com ([119.145.14.66]:31737) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZWa3q-0007V1-ED for qemu-devel@nongnu.org; Mon, 31 Aug 2015 21:08:30 -0400 References: <1438159544-6224-1-git-send-email-zhang.zhanghailiang@huawei.com> <1438159544-6224-20-git-send-email-zhang.zhanghailiang@huawei.com> <55E0DCFB.2070402@redhat.com> <55E41E1D.7020500@huawei.com> <55E46D9B.2040908@redhat.com> From: zhanghailiang Message-ID: <55E4FA7B.7080305@huawei.com> Date: Tue, 1 Sep 2015 09:08:11 +0800 MIME-Version: 1.0 In-Reply-To: <55E46D9B.2040908@redhat.com> Content-Type: text/plain; charset="utf-8"; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH COLO-Frame v8 19/34] qmp event: Add event notification for COLO error List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eric Blake , qemu-devel@nongnu.org Cc: lizhijian@cn.fujitsu.com, quintela@redhat.com, Markus Armbruster , yunhong.jiang@intel.com, eddie.dong@intel.com, peter.huangpeng@huawei.com, dgilbert@redhat.com, arei.gonglei@huawei.com, amit.shah@redhat.com, Michael Roth On 2015/8/31 23:07, Eric Blake wrote: > On 08/31/2015 03:27 AM, zhanghailiang wrote: > >>>> +Data: None. >>>> + >>>> + - "mode": COLO mode, 'primary' or 'secondary' >>>> + - "error": Error message (json-string, optional) > > The "Data: None" designation is inconsistent with the fact that you do > provide data. > Will fix in next version. >>>> + >>>> +Example: >>>> + >>>> +{"timestamp": {"seconds": 2032141960, "microseconds": 417172}, >>>> + "event": "COLO_EXIT", "data": {"mode": "primary"}} >>> >>> It might also be useful to provide a machine-parseable parameter on >>> whether the exit was due to an internal error vs. an external request. >>> Maybe by adding 'flag':'bool', since presence or absence of 'error' is >>> not as nice. >>> >> >> Good idea, but i think the type of 'flag' changes to 'str' maybe better, >> which can be 'error' or >> 'request', 'bool' is not so self-explanatory. >> The original reason for adding 'error' member is to help >> users to know what exactly happen, but it seems to be difficult to >> exactly class all errors, >> so i will remove it. > > Keeping an optional 'error' for human consumption is not bad; it's just > that it is also nice to provide something for machine consumption (and > we typically document that machines should NOT parse 'error' messages). > OK, then i will keep it, and manage all the error places and error messages carefully. Thanks, zhanghailiang