On Mon, 2007-09-10 at 02:23 -0400, Bob S wrote: > On Sunday 09 September 2007 14:30, Peter Sjoberg wrote: > > > On Sat, 2007-09-08 at 00:49 -0400, Bob S wrote: > > > Hello SuSE people, > > .<snip a bunch>................ > > > I love lvm since it's so flexible, if you for example run low in space > > in datalv you can just expand it without playing around with disk > > partitions (=much safer) and you can even add a new disk and expand it > > without problem. If I need to replace a disk with a bigger one/remove > > one I can use a single "pvmove /dev/hdb" to move data around and get it > > done without tons of repartition and fs moves. > > OK, but I have a question; > > When I installed 10.0 I used LVM for everything except /swap and /. When I > went to install 10.2 I was going to use LVM again, but when I looked at the > partitioner it seemed to want put my 10.2 partitions with the old 10.0 stuff > under /system2, That worried me. How could I have a homelv for 10.0 and a > homelv for 10.2? Well, that's one thing you need to keep track of since you can't have two lvms with the same name. If you want homelv for 10.0 and 10.2 you can but you need to name them something like home0lv and home2lv. Same goes for anything else that has naming conflict. The system2 vg can be common between all of them.
> (relying on memory here which isn't as good as it used to > be) How would the 10.0 os and the 10.2 os sort it out. Soooo, I just resorted > to a regular partitioning scheme for 10.2 It's a bit like having a disk called system2 and partitions called homelv/homel2lv/fc5varlv/fc5usrlv... If it makes you feel more comfy, name them sdb1,sdb2,sdb3 :) > > > > One thing is that since /boot and /boot/grub/menu.lst is common for all > > installs you need to manually manage that area. I found that each os > > version have there own version numbering like > > vmlinuz-2.6.18.8-0.5-default/initrd-2.6.18.8-0.5-default so there is no > > conflict > > Ummm,,, that would be the kernel version, so it wouldn't be right if the > kernel were upgraded, right? When I upgraded the kernel in 10.2 I lost the > ability to boot 10.0. I attributed it to that but I didn't look into it > because 10.2 was working well. Guess I would have to reinstall 10.0 or find > the proper kernel and install it. I don't like the way suse updater just replaces the kernel without leaving the previous as a fallback (caused huge issue when I upgraded a system that needed customized network drivers and I couldn't download them for the new version since I didn't had a working network). On the good news side, at most it may remove the old version from menu.lst but it will add the new version somewhere so when you boot you can go up/down in the menu to entry that works. > > > but they normally replace /boot/grub/menu.lst so I make sure I > > have a copy of menu.lst somewhere and then I manually merge the old and > > new menu.lst after each install. > > So you must have both kernel versions on your system. /boot contains each os versions (all) kernel/initrd version. > I don't like the > automated update of the kernel because it replaces it rather than adding the > new kernel and you end up with only the one kernel. Agree > Way back when.... I would > download a kernel and manually install it. That way I would have more than > one kernel to fall back on. When you have several different os versions around you have a complete different os to fall back on, better then a rescue cd in many cases. > I guess I will start keeping a copy of Menu.1st > also. > > > > > Anxiously awaiting the final 10.3 so I can try Compiz-Fusion, Beryl > > > whatever and be able to fall back to 10.2 when I screw it up. > > > > I'm also waiting for 10.3 final but you can do as me and start playing > > with Beta 3 to get a feel for it and report problems (or you may have to > > report same problem on the final because everyone assumed someone else > > already tested and reported it) > > Welllll.....I will, as soon as I free up the space on my drive which I > screwed > up by improperly partitioning it. Thanks for your input. I did see your "corrupt" config with the extended partition not covering the rest of the disk. I don't know if it's any safe way to change end cylinder of sda4 to 30401 to make the disk usable again (Hmm, maybe "sfdisk -d", need to test that in my lab in case I hit it). If all else fails, one option would be to drop sda3 (swap) and create sda3 as all the free space (and make it lvm). You still "loose" 2G but right now you lost a lot more. > > Bob S. /ps -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
