I have increasingly seen situations where a user has a default fedora installation (say FC6) and they decide to upgrade, and upgrade the HDD at the same time.
eg.
http://www.linuxquestions.org/questi....php?p=2825663
http://www.linuxquestions.org/questi....php?p=2798327
They now have their brand new default f7 installation, plug in the old drive and expect to be able to mount it's ext3 partitions and just copy their data over. This would seem a reasonable expectation... only the defaults they have trusted have made an LVM volume out of their single drive. They get an error:
Code:
mount: wrong fs type, bad option, bad superblock on /dev/sdb2,
missing codepage or other error
The good ones do the right things... they check /etc/fstab and fdisk -l and discover the LVM thing they've never heard of.
This is the point we expect users to issue plaintiff cries... IMO, justified: they're not expecting this!
However: this does not happen. Instead they do the correct, non-luser, thing and google about LVM. They read the
man page and find a huge wealth of information.
http://www.tldp.org/HOWTO/LVM-HOWTO/index.html
All of it technical.
There is confusion because their old and new volumes have the same label.
Under a calm guidance, and since these are clearly not the run-of-the-will user, things get resolved. But I would have thought it would be simpler than this - handled at a higher level.
I can even sort-of see how... where a new volume (with an older datestamp) is introduced, which has the same VG name (i.e. VolGroup01) as an existing one, it seems this could well be a previous installation and could be common enough an occurrance to be anticipated.
In short: surely there is a standard tool for dealing with this?
I haven,t found one... but I am a simple soul.
I don't have a fedora install here to experiment with either. It isn't something I have had to deal with but for others: I haven't used a default install since RH9. I normally advise (and use) a custom scheme without LVM and a separate home partition for fedora.
Instead there is the old vgscan followed by vgchange -a y ... which is not all that worse than the regular mount command, just unexpected.
Can anybody shed light on this?
The purpose of this thread is part rant (so you good folks can tell me how silly I'm being) and part "collecting information" so a group of users who have been identified actually
using the search functions will have some place to land.
No, I don't think this is bad enough to call a "bug" (in implementation) but if I get much more of these I'll reconsider.