Received-SPF: pass (google.com: domain of linux-nfs+bounces-2207-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
From: Benjamin Coddington <bcodding@redhat.com>
To: Christian Theune <ct@flyingcircus.io>
Cc: Chuck Lever <chuck.lever@oracle.com>, linux-nfs@vger.kernel.org,
 jlayton@kernel.org, linux-kernel@vger.kernel.org,
 Linux regressions mailing list <regressions@lists.linux.dev>,
 trondmy@hammerspace.com, anna@kernel.org
Subject: Re: nfs4_schedule_state_manager stuck in tight loop
Date: Tue, 05 Mar 2024 06:39:25 -0500
Message-ID: <9FB1C80C-CF8A-4650-8918-475A19F353EB@redhat.com>
In-Reply-To: <3404A3C2-31BE-4D39-A256-E3A1FB48ACFB@flyingcircus.io>
References: <8B04DA70-ABEF-44A4-BBA7-60968E6CFA10@flyingcircus.io>
 <ZdisssP88_9o0BXn@manet.1015granger.net>
 <3404A3C2-31BE-4D39-A256-E3A1FB48ACFB@flyingcircus.io>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

On 5 Mar 2024, at 1:09, Christian Theune wrote:

> Hi,
>
> not sure whether I may have missed a response that didn=E2=80=99t make =
it back to me or any of the lists.
>
> Just in case, because the CC didn=E2=80=99t include the original addend=
um I made to my report:
>
> Addendum:
>
> I=E2=80=99ve checked kernel changelogs since then but didn=E2=80=99t fi=
nd anything that I could relate to this aside from *maybe* dfda2a5eb66a68=
5aa6d0b81c0cef1cf8bfe0b3c4 (rename(): fix the locking of subdirectories) =
which mentions NFS but doesn=E2=80=99t describe the potential impact.
>
> We=E2=80=99re running 5.15.148 now and as it=E2=80=99s been another 2 m=
onths there might be the chance of another lockup in the near future ;)
>
> If anyone has ideas on how to debug/approach a reproducer I=E2=80=99d b=
e more than happy to help and try to provide more data.
>
> Cheers,
> Christian

When the problem occurs, use the

nfs4:nfs4_state_mgr
nfs4:nfs4_state_mgr_failed

tracepoints to see what the state manager might be doing.

Also a network capture might show what the state manager thread is up to =
if
it is sending operations.

Ben