Comment 3 for bug 496796

pitti, I already had looked at this with sarvatt for a while yesterday. While it results in the user ending up in failsafeX, it's not a bug in failsafeX but rather in whatever is causing it to fall into failsafeX. I gathered from the logs sarvatt had collected (which don't appear to be attached to this bug report - sarvatt you should upload them here) that gdm was trying to launch X on tty7 which was being held open by whatever was running the fsck so gdm gave up. failsafeX is less picky about what vt it fires up on which is why it was able to come up successfully.

My guess is that some sort of magic in the gdm upstart script (or even prior to this) is required in order to serialize startup when fsck is proceeding. For instance, gdm maybe should wait and retry if tty7 is temporarily unavailable before trying to start X on it, or else allow gdm to fire up X on a different tty if you really do actually want X to come up while fsck is proceeding in the background.