On Thu, Apr 20, 2017 at 11:12 AM, Niels Thykier <[email protected]> wrote: > Control: tags -1 moreinfo > > On Mon, 11 Jul 2016 21:05:25 -0300 Felipe Sateler <[email protected]> > wrote: >> On Mon, 04 Jul 2016 22:27:20 -0400 "Roberto C. Sanchez" >> <[email protected]> wrote: >> > Package: lintian >> > Version: 2.5.45 >> > Severity: normal >> > >> > -----BEGIN PGP SIGNED MESSAGE----- >> > Hash: SHA512 >> > >> > In preparing a new upstream release of the shorewall-init package, I >> > found that lintian gave the following errors: >> > >> > E: shorewall-init: init.d-script-missing-dependency-on-remote_fs >> > etc/init.d/shorewall-init: required-start >> > E: shorewall-init: init.d-script-missing-dependency-on-remote_fs >> > etc/init.d/shorewall-init: required-stop >> > >> > I checked with upstream was informed that shorewall-init doesn't work >> > with NFS mount /usr and so assumes /usr is accessible without the need >> > for $remote_fs in Required-Start and Required-Stop. I inquired in >> > #debian-devel on OFTC about this to see if it was OK to override. I was >> > told that this error is spurious as initramfs now mounts /usr and was >> > suggested to file a bug report against lintian. >> > >> >> Moreover, booting without /usr mounted is now officially[1] >> unsupported, so there is no point in playing /-/usr whack-a-mole. >> >> >> [1] #830829 >> >> Saludos >> >> > > Hi, > > Should we remap this as a check for "$local_fs" or can people assume > that /usr is always present?
I'd say yes. The idea is that it should be guaranteed that /usr is mounted by the time the initramfs passes control to the init system. This means that most software (except things that survive past the init killing spree) can assume /usr is always available. > Currently, use of /var needs "$local_fs". I'd say that use of /var requires $remote_fs. Remote-mounted /var is a supported configuration. -- Saludos, Felipe Sateler
