[ovs-dev] [PATCH] tests-windows: Fix SSL ovsdb test which is hanging

Ben Pfaff blp at ovn.org
Thu Mar 8 01:07:08 UTC 2018


On Thu, Mar 08, 2018 at 02:22:45AM +0200, aserdean at ovn.org wrote:
> > -----Mesaj original-----
> > De la: ovs-dev-bounces at openvswitch.org <ovs-dev-
> > bounces at openvswitch.org> În numele Ben Pfaff
> > Trimis: Thursday, March 8, 2018 2:11 AM
> > Către: aserdean at ovn.org
> > Cc: dev at openvswitch.org
> > Subiect: Re: [ovs-dev] [PATCH] tests-windows: Fix SSL ovsdb test which is
> > hanging
> > 
> > On Thu, Mar 08, 2018 at 01:45:51AM +0200, aserdean at ovn.org wrote:
> > > > -----Mesaj original-----
> > > > De la: ovs-dev-bounces at openvswitch.org <ovs-dev-
> > > > bounces at openvswitch.org> În numele Ben Pfaff
> > > > Trimis: Wednesday, March 7, 2018 11:20 PM
> > > > Către: Alin Gabriel Serdean <aserdean at ovn.org>
> > > > Cc: dev at openvswitch.org
> > > > Subiect: Re: [ovs-dev] [PATCH] tests-windows: Fix SSL ovsdb test
> > > > which is hanging
> > > >
> > > > On Tue, Mar 06, 2018 at 01:17:12PM +0200, Alin Gabriel Serdean wrote:
> > > > > The test:
> > > > > `1827. ovsdb-server.at:490: testing SSL db: implementation ...` is
> > > > > hanging on Windows because the returned in the case the client
> > > > > failed to connect is "Unknown error" vs the normal "Protocol error".
> > > > >
> > > > > Update the test to accommodate for this.
> > > > >
> > > > > Signed-off-by: Alin Gabriel Serdean <aserdean at ovn.org>
> > > >
> > > > Thanks for finding and fixing the error.
> > > >
> > > > I think that there is an underlying problem here: the code is bad at
> > > > error handling in general.  It's supposed to kill the ovsdb-server
> > > > if anything fails, but it didn't do it right: it killed $(cat pid)
> > > > but the pidfile is actually in ovsdb- server.pid.  This meant that
> > > > the issue showed up as a hang instead of a test failure, which seems bad.
> > > >
> > > > Also, I'm kind of inclined to just disregard the particular error
> > > > message.  It might save time later.  (I guess there's a risk that
> > > > the connection fails happens for some other reason that should fail
> > > > the test
> > > > though.)
> > > >
> > > > Anyway, I sent a pair of patches that implement my suggestions:
> > > > https://patchwork.ozlabs.org/patch/882796/
> > > > https://patchwork.ozlabs.org/patch/882798/
> > > >
> > > > What are your thoughts?
> > > >
> > > > Thanks,
> > > >
> > > > Ben.
> > > [Alin Serdean] I'm dropping this change and acking yours.
> > > They are much cleaner and better 😊.
> > 
> > Thanks!  I applied both of these fixes to master.
> > 
> > Do you want me to backport them to branch-2.9 (or earlier)?  I have not done
> > that yet.
> 
> branch-2.9 also please.
> 

Done.


More information about the dev mailing list