Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Sun, 23 May 2021 21:14:37 +0200
From:   Petr Vorel <pvorel@suse.cz>
To:     "J. Bruce Fields" <bfields@redhat.com>
Cc:     linux-nfs@vger.kernel.org, Yong Sun <yosun@suse.com>
Subject: Re: pynfs: COUR2, EID50 test failures
Message-ID: <YKqpnXjGP+Ub0lB0@pevik>
Reply-To: Petr Vorel <pvorel@suse.cz>
References: <YKdXZa6Izs7kqgfE@pevik>
 <YKf+a7e4lv2BwBBQ@pick.fieldses.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YKf+a7e4lv2BwBBQ@pick.fieldses.org>
Precedence: bulk

> On Fri, May 21, 2021 at 08:47:01AM +0200, Petr Vorel wrote:
> > Hi Bruce,

> > what's wrong with pynfs COUR2, EID50 tests?
> > They're failing on various kernels.
> > I got these failures on 5.11.12 (openSUSE package), 5.13.0-rc2 (mainline on
> > openSUSE), 4.9.0 (Debian package):

> > COUR2    st_courtesy.testLockSleepLock                            : FAILURE
> >            OP_OPEN should return NFS4_OK, instead got
> >            NFS4ERR_GRACE

> Are you running it immediately after starting the server?  If so, that's
> expected.  Personally my test scripts mount the server and create a file
> there before running pynfs tests.  The create won't return until the
> grace period's done.
Thanks for info, I'll test that!

> Might be worth documenting in the README, or teaching pynfs to wait
> before running tests.
Or maybe test should create a test file before? That might be faster than waiting.
Because there is some wait for 2 min as part of testing.

> > EID50    st_exchange_id.testSSV                                   : FAILURE
> >            nfs4lib.NFS4Error: OP_EXCHANGE_ID should return
> >            NFS4_OK, instead got NFS4ERR_ENCR_ALG_UNSUPP

> > Nothing in dmesg (tested with "rpcdebug -m nfsd -s lockd").
> > Or is it just my wrong setup?

> SSV support is optional.  We should probably drop the "all" tag from
> that test.
Again, thanks for info. It'd be great if you implement that.

Kind regards,
Petr

> --b.