From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2) Gecko/20021126 Description of problem: See bug 77539 - which is closed as "notabug" which I believe is wrong. When you take a system with no LVM partitions and add one (say /fred) and add that to the fstab, when you reboot the system the fsck on the filesystem with fail because the LVM initialisation didn't happen because the LVM module is not loaded. Previous reports of this bug have been closed because "you need to include the LVM in the initrd" ... which is fine, but it isn't intuitive and mkinitrd won't include the module by default unless the root filesystem is on a LV (which means each time I upgrade my kernel I will break my machine again!) There are several other modprobes in rc.sysinit, and additional LVM one wouldn't hurt (if /etc/lvmtab exists etc). Or, if this really can't be done, then this problem (and workaround) should be clearly documented where people are going to find it when they start playing with LVM. It's small useability inconsistancies like this that scare people away from Linux. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Build a non-LVM system 2. Create a LV based filesystem and add it to fstab 3. Reboot Actual Results: fsck fails because LVM initialisation is not done. Expected Results: No fsck problems (and correct LVM initialisation) Additional info:
This is tweaked in the current initscripts in rawhide; as long as an lvmtab exists, the module will be loaded.