Re: [10.0_STABLE] Hard lock

[Michael van Elst <mlelstv%serpens.de@localhost>]

On Fri, Nov 22, 2024 at 12:42:59PM +0100, BERTRAND Joël wrote:
> Michael van Elst a écrit :
> > On Thu, Nov 21, 2024 at 08:26:11PM +0100, BERTRAND Joël wrote:
> > Except for the wild mixup of early kernel messages (at time ~2.204)
> > with iscsi debug messages (at ~77.724) this only hints to a cleanup
> > of a session.
> 
> 	I understand. But how can this session be cleaned up ?
> iSCSI target is used and fsck has just started.

When the network fails (e.g. timeout), a connection will be closed.
The code tries to reconnect automatically, but after some retries
it gives up. When it was the last connection (usually there is only
one per session), then the session is closed and the disk device
is detached from the system.

We have already seen some network problem and it may even cause this.
But the result should be a failure when accessing the disk. If
that's fsck, it should report some I/O error. In no way should it
lock up the system.


Greetings,
-- 
                                Michael van Elst
Internet: mlelstv%serpens.de@localhost
                                "A potential Snark may lurk in every tree."