[ovs-dev] [PATCH] tests/ovs-monitor-ipsec: Remove --no-wait -vreconnect:emer from ovs-vsctl.

[Ben Pfaff]

On Fri, Jul 22, 2016 at 11:02:02AM -0700, Joe Stringer wrote:
> On 21 July 2016 at 18:55, Daniele Di Proietto <diproiettod at vmware.com> wrote:
> > This fixes a sporadic failure of the testcase, where ovsdb-server.log
> > contains:
> >
> > 2016-07-21T21:08:22.279Z|00003|jsonrpc|WARN|unix: send error: Broken pipe
> > 2016-07-21T21:08:22.279Z|00004|reconnect|WARN|unix: connection dropped (Broken pipe)
> >
> > I'm not sure what was the purpose of the original "--no-wait
> > -vreconnect:emer", but the test seem to run fine without it.
> >
> > Signed-off-by: Daniele Di Proietto <diproiettod at vmware.com>
> 
> This seems like it was a common technique used in a lot of the older
> tests to try to modify the database if not all the daemons are
> running. In practice, these days OVS_VSWITCHD_START provides
> assurances for this, so it should be unnecessary (indeed, few tests
> still do this). The -vreconnect:emer should hide warnings like the
> above, which I suspect it does for the ovs-vsctl invocation, however
> this doesn't prevent the final OVS_VSWITCHD_STOP from detecting this
> log and complaining.
> 
> Acked-by: Joe Stringer <joe at ovn.org>

Should we audit other tests for this?