[rhinstaller/anaconda/pulls/209 master] fix storage writing for live and ostree installs (#1236937)

AdamWill installerbot-noreply at redhat.com
Thu Jul 16 19:00:53 UTC 2015


OK, how about this version? It does what I suggested in the last comment. Locally I added a couple of debug messages right above the `self.storage.write()` calls so I could tell for sure when the storage write happened; I then tested a regular live install and a boot.iso install. As intended, only `writeStorageEarly()` ran on the boot.iso install and only `writeStorageLate()` on the live install. Both installed systems had a `/etc/fstab` written by anaconda once install was complete.

I also thought of an even simpler alternative to this: simply run both `Early` and `Late` for all installs. AFAICS, we have `Early` because for a package-based install, some package scripts need `/etc/fstab` to exist, so we want to write it before the package install process starts...but there would be no harm caused by writing it *again* after the install process was complete. We have `Late` because on live-style installs where we dump an entire disk image to the system that overwrites the `/etc/fstab` written by `Early`...but again, there'd be no harm caused by writing it Early as well as Late in that case. So, wouldn't it actually be fine to just write it twice for all installs? (well, we can keep the `if not flags.dirInstall` checks and not do storage writing at all when `flags.dirInstall` is set, I presume that's there for a good reason).
-- 
To view this pull request on github, visit https://github.com/rhinstaller/anaconda/pull/209


More information about the anaconda-patches mailing list