Content-Type: text/plain;
        charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 11.5 \(3445.9.1\))
Subject: Re: [PATCH] lockd: NSMPROC_MON should be send only once even if in
 multithread
From:   Chuck Lever <chucklever@gmail.com>
In-Reply-To: <bdef1ea6-1b71-a638-7274-d2109db5e1d1@huawei.com>
Date:   Thu, 24 Jan 2019 11:17:48 -0800
Cc:     Bruce Fields <bfields@fieldses.org>,
        Jeff Layton <jlayton@kernel.org>,
        Trond Myklebust <trond.myklebust@hammerspace.com>,
        Anna Schumaker <anna.schumaker@netapp.com>,
        Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <DE9790AD-A95E-43CB-9A5C-A55DD10294CC@gmail.com>
References: <1547705746-69554-1-git-send-email-zhangxiaoxu5@huawei.com>
 <4E812155-4DB4-4EF4-91ED-EDBB1B0BBFF7@gmail.com>
 <00390d85-d2dd-be57-4192-bff562a75684@huawei.com>
 <E87ADB64-FE89-48A8-A401-3153E029D0DA@gmail.com>
 <bdef1ea6-1b71-a638-7274-d2109db5e1d1@huawei.com>
To:     "zhangxiaoxu (A)" <zhangxiaoxu5@huawei.com>
Sender: linux-nfs-owner@vger.kernel.org
Precedence: bulk


> On Jan 24, 2019, at 4:29 AM, zhangxiaoxu (A) <zhangxiaoxu5@huawei.com> =
wrote:
>=20
>=20
>=20
> On 1/19/2019 4:59 AM, Chuck Lever wrote:
>>> On Jan 17, 2019, at 9:48 PM, zhangxiaoxu (A) =
<zhangxiaoxu5@huawei.com> wrote:
>>>=20
>>>=20
>>>=20
>>> On 1/17/2019 10:33 PM, Chuck Lever wrote:
>>>> Hi, what is the harm of sending more than one MON request for a =
peer?
>>> Maybe no harm.
>>> The rpc.statd won't record the peer twice.
>>>=20
>>> I found this when I tested the xfstest generic/089.
>>> The rpc task for that msg sometimes take very long time.
>>> rpc took 57 sec who t_mtab/2377 srv rpc.statd xid 2453489031 prog =
statd/100024/1 proc 2 prot 6 flags 0x680
>> MON is supposed to be a call to a local service (on the same host).
>> It would be interesting if you could determine why it takes so long.
> Yes, I'm working for that.
> I think this is an optimization point.
> Why do we have to send multiple requests?

I agree that multiple requests are unnecessary, but
they are harmless too. I don't think it's worth the
trouble to change this behavior.

The problem is MON requests on your system takes too long.


>>> I think the msg just send only one time is enough.
>>> RPC and rpc.statd maybe consume for a long time on the network.
>> --
>> Chuck Lever
>> chucklever@gmail.com

--
Chuck Lever
chucklever@gmail.com