I noticed that blivet in updateSizeInfo() always makes a full e2fsck on all detected partitions, when running anaconda in rescue mode, and selecting 1/ --continue-- , which is somewhat an unwanted behavior, as some fsck operations can take a *very* long time on some systems. The only way to bypass this forced fsck is to drop to a shell immediately, and to manually detect, setup and mount all the needed partitions. This behavior changed since the F-23 installer. Maybe blivet could be put in installer_mode also when anaconda is running is rescue mode, which I think is not the case ?
Err, blivet *is* put in installer_mode actually (blivet/formats/fs.py): > if flags.installer_mode and self._resize.available: > # if you want current/min size you have to call updateSizeInfo > try: > self.updateSizeInfo() > except FSError: > log.warning("%s filesystem on %s needs repair", self.type, self.device)
This also happens in Fedora 25 when clicking the "Continue" button in the installer after selecting the language. The installer appears to hang during the fsck. 15 minutes so far...
*** Bug 1189905 has been marked as a duplicate of this bug. ***
*** This bug has been marked as a duplicate of bug 1170803 ***