On Fri, Jan 12, 2018 at 11:18:01PM +0100, Martin Wilck wrote: > On Fri, 2018-01-12 at 22:47 +0100, Julian Andres Klode wrote: > > > > I don't know actually, I'm just updating the package :) > > The feature needs to be enabled explicitly, so indeed you weren't using > it. > > > >I can say that > > the tests were run with default configuration, normally we run with > > DEFAULT_FIND_MULTIPATHS but that does not work yet with 0.7.4 - no > > dm-0 device seems to get created in this test case (some systems > > without > > multipath were broken 2.5 years ago just by installing it because it > > took over non-multipath devices, hence the option was enabled; might > > not even be needed anymore). > > That's still the default behavior if find_multipaths is off. And > actually, I don't know what's wrong with that. If you don't like it, > blacklist your root device. All you need to make sure is that multipath > settings in the initrd and in the booted system are the same. > > "find_multipaths" is not a silver bullet to kill off multipath-related > boot problems. It has advantages and disadvantages, depending on the > use case. Anyway, I'm not aware of general problems with it in the > later multipath versions. I'll try to give it a bisect since the test works in 0.6.4 and not in 0.7.4. It's going to be painful :) > > Btw, one thing you should be aware of is that udev rule files and > systemd service files have changed. That requires also dracut changes > that are not in upstream dracut yet. We don't use dracut, we use initramfs-tools, but yes, I have to update the hooks for that too. -- debian developer - deb.li/jak | jak-linux.org - free software dev ubuntu core developer i speak de, en -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
