On Wed, Apr 04, 2018 at 02:22:25PM +0200, Timothy Redaelli wrote: > On Tue, 3 Apr 2018 11:06:02 -0700 > Ben Pfaff <[email protected]> wrote: > > > On Tue, Apr 03, 2018 at 03:38:28PM +0200, Timothy Redaelli wrote: > > > Currently some tests fails if run with multiple jobs at once, so > > > this commit disables parallel jobs during the recheck phase. > > > > > > Signed-off-by: Timothy Redaelli <[email protected]> > > > > I think that this is a useful patch. > > > > To clarify, though, do you see some tests that always fail when run in > > parallel? That would be new to me, and I'd place a higher priority on > > fixing it. I do see that some tests are more likely to fail that way. > > Some tests fails often than other tests (when run in multiple jobs), so > rarely (on a big machine, with 32 cores and by using -j32) a test > fails twice and so make check RECHECK=yes fails. > This patch only avoid this rare failure that can be annoying if you are > releasing a new package (on Fedora or RHEL for example) since you need > to re-build all the architectures only for a false positive on a > single architecture.
I've seen that in Debian, too. It has more architectures so it can be worse. Thanks for clarifying. I added that information to the commit message and applied this to master. If it would help to backport to branch-2.9, let me know. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
