Re: [Qemu-devel] [PATCH v7 2/3] qmp event: Refactor QUORUM_REPORT_BAD

[Eric Blake <eblake@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1640 bytes --]

On 02/25/2016 06:39 PM, Changlong Xie wrote:
> Introduce QuorumOpType, and make QUORUM_REPORT_BAD compatible
> with it.
> 
> Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
> Cc: Wen Congyang <wency@cn.fujitsu.com>
> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
> Signed-off-by: Changlong Xie <xiecl.fnst@cn.fujitsu.com>
> ---

> +Read operation:
>  { "event": "QUORUM_REPORT_BAD",
> -     "data": { "node-name": "node0", "sector-num": 345435, "sectors-count": 5 },
> +     "data": { "node-name": "node0", "sector-num": 345435, "sectors-count": 5,
> +               "type": "read" },
>       "timestamp": { "seconds": 1344522075, "microseconds": 745528 } }
>  
> +Flush operation:
> +{ "event": "QUORUM_REPORT_BAD",
> +     "data": { "node-name": "node0", "sector-num": 0, "sectors-count": 2097120,
> +               "type": "flush", "error": "Broken pipe" },
> +     "timestamp": { "seconds": 1456406829, "microseconds": 291763 } }
> +
>  Note: this event is rate-limited.

Question - do we care if rate limiting masks one type of failure due to
another?  Or put another way, are we okay with a single rate-limiting
queue for all three types, or do we want three queues?  Also, shouldn't
this have a queue per child node (I don't want to be flooded with
multiple notifications in one second that child1 has failed, but I _do_
want notifications if both child1 and child2 fail in the same second).

But that's for future patches to change; it does not need to hold up the
current series.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 604 bytes --]