![](https://seccdn.libravatar.org/avatar/7f15e26f6f6f56d85aa0ec0fc4e37c44.jpg?s=120&d=mm&r=g)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 jdd wrote:
Pascal Bleser wrote: ...
automatically propose the default partitioning setup with LVM. I never could know what, with LVM, happens of the system file if _one_ disk (or partition) is destroyed. is not the entire file system destroyed?
Actually, when you have several (physical) harddisks in an LVM Volume Group (which can be seen as a virtual disk) and one of them crashes, the whole Volume Group is corrupted on the blocks that were stored on the broken harddisk. I never had the case (and on critical/production systems, I use it on top of software or hardware RAID anyway, see below), so I don't know how LVM behaves exactly, but it's pretty obvious from the way LVM works. To avoid that, you must use LVM on top of RAID (e.g. mirroring), which is feasible too (and works pretty well), although there was a little bugger in YaST2's partitioner with previous releases (don't know about 10.0) where you couldn't create LVM on top of RAID on one pass: one had to first create the RAID partitions, then reboot and start the installation again, then create the LVM on top of the RAID partitions. As I said, I didn't check that on 10.0 (and I don't have spare partitions to test that on, at the moment). When using LVM on top of RAID (not RAID 0, obviously), the RAID subsystem makes sure data is mirrored and when one disk dies, etc etc.. And LVM builds its Volume Groups and Logical Volumes on top of the RAID, which means that when one physical disk dies, you don't loose any data. - -- -o) Pascal Bleser http://linux01.gwdg.de/~pbleser/ /\\ <pascal.bleser@skynet.be> <guru@unixtech.be> _\_v The more things change, the more they stay insane. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) iD4DBQFDECbWr3NMWliFcXcRAkiEAJ45hL6KMxBO2yzK3/5qIVL6RCc0BgCY8Lep ahO0KTbP0F8YMxSKp33s0w== =HvLK -----END PGP SIGNATURE-----