Re: migration-test random intermittent failure, openbsd VM

[Daniel P . Berrangé]

On Thu, Feb 08, 2024 at 03:04:23PM -0300, Fabiano Rosas wrote:
> Peter Maydell <peter.maydell@linaro.org> writes:
> 
> +cc Daniel.
> 
> > Random intermittent in migration-test when running the tests
> > in the openbsd VM (i.e. what you get from 'make -C build vm-build-openbsd')
> > Any ideas?
> 
> Where's your HEAD at?
> 
> >
> > 106/847 qemu:qtest+qtest-x86_64 / qtest-x86_64/migration-test
> >            ERROR          157.51s   killed by signal 6 SIGABRT
> > ――――――――――――――――――――――――――――――――――――― ✀  
> > ―――――――――――――――――――――――――――――――――――――
> > stderr:
> > qemu-system-x86_64: multifd_send_sync_main: channel 12 has already quit
> > qemu-system-x86_64: TLS handshake failed: The TLS connection was
> > non-properly terminated.
> > qemu-system-x86_64: TLS handshake failed: Error in the push function.
> > qemu-system-x86_64: Failed to connect to '127.0.0.1:23083': Address
> > already in use
> 
> Looks like a legitimate port clash here when doing the TLS
> connection. It seems the test is not prepared to deal with it.

The migration-test.c code always says to listen on ':0' so the
kernel dynamically allocates an address, we when query that
port and connect to it.  So we should never fail on listen/bind
server side.

The error message though is about the client side as it says
"failed to connect".

I've no idea how on earth you trigger "Address already in use" as
a client !

Every outbound connection does require a port, but QEMU always
leave the port on 0, so the kenrel dynamically allocates an
outbound port. Somehow this is failing, but I can't see how
it is QEMU's fault ?!?!?


With regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|