[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <qbzzr6waj23q6rxrueevwoteg5i4ogr7hr45ckseu7ekdcc7sb@xgyvtkp27w5i>
Date: Thu, 7 Mar 2024 13:43:33 +0100
From: Daniel Wagner <dwagner@...e.de>
To: Sagi Grimberg <sagi@...mberg.me>
Cc: Hannes Reinecke <hare@...e.de>, James Smart <james.smart@...adcom.com>,
Keith Busch <kbusch@...nel.org>, Christoph Hellwig <hch@....de>, linux-nvme@...ts.infradead.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 2/5] nvme-fc: do not retry when auth fails or
connection is refused
On Thu, Mar 07, 2024 at 12:25:16PM +0200, Sagi Grimberg wrote:
> > Is this what you had in mind?
> >
> > Which, incidentally, is basically the patch I just posted.
>
> Reading this, the patchset from Hannes now is clearer.
> Isn't the main issue is that nvme-fc tries to periodicly reconnect
> when failing the first connect? This is exactly what the test expects
> it to do right?
Yes, the test expects that the initial connect attempt fails. nvme-fc is
using one connect path and doesn't distinguish between the initial
connect attempt and a reconnect.
All fabric transport share the same problem when the connection has been
established and later one a connection drop happens and a reconnnect is
executed. The blktest nvme/048 case extension I've posted tests the
reconnect attempt after the controller key has changed. In this
scenario, nvme-tcp, nvme-rdma will also do a reconnect attempt although
DNR is set.
Powered by blists - more mailing lists