ext2 seems kinda dodgy - what is best alternative with a laptop?
Hi - i have a laptop with slackware 10.1-2.6.15
currently the drive has various partitions including Fat32/NTFS and ext2 for linux. the ext2 system seems pretty dodge in my opinion (files seem to somestimes disappear and if my machine ever crashes or has to get hard power-cycled or runs out of battery all sorts of badness seems to crop up....) It seems i can either 'convert' to ext3 or 'backup/delete' existing partition and then re-create as reiserfs. My questions are: 1) Is ext3 slower/same/faster than ext2? 2) is reiserfs better than ext3 (if so why?) 3) if i was to opt for reiserfs and created a new partition, can i restore my backed up ext2 data onto this volume? 4) are they're any downsides to either ext3 or reiserfs when using a laptop computer (ie. it runs off batteries - do these filesystems make more frequent writes to the drive compared to ext2? ) Are there any recomended tools for this task? I have the Linux SystemRescue boot CD which seems to have a whole range of goodies that have been very usefull to date. |
"Better" is a very subjective term. ext2 is "dodgy" in that it doesn't have a journal which all modern filesystems do. It'll treat you right if you treat it right and hard resets etc aren't very nice. I use ext3 because it works exceptionally well in my opinion. I spend a lot of time on these forums and I regularly see people asking how to recover their reiser partition that crapped out on them but I never see people having problems with ext3. That's why I use it and it's given me everything I want the last few years :)
|
Turning an ext2 filesystem into a ext3 filesystem can be done live, on the fly:
Code:
tune2fs -j /dev/<your_ext2_partition> Eric |
You should know that it will be nearly impossible to recover deleted files from journaling filesystem, which is possible using ext2.
|
Quote:
|
As I know, none of the journaling filesystems can be recovered.
|
Cs-Cam / Alien Bob / Alien_Hominid
Love your work guys!! :) thanks for the cripsy clear responses... I will be upgrading this puppy to ext3 filesystem tonight. If there is chance of non recoverable data with journaling fs it looks like ext3 is the more reliable of the 2. Entire partition loss is not such a big deal for me at the moment anyway as i always dump a clean image after a new build. |
well its been nearly 3 minutes and i'm back again... that all happened a little quicker than i thought it would! LOL...
root@Slack700m:~# tune2fs -j /dev/hda6 tune2fs 1.35 (28-Feb-2004) Creating journal inode: done This filesystem will be automatically checked every 37 mounts or 180 days, whichever comes first. Use tune2fs -c or -i to override. edited /etc/fstab is that it ? hehe... im used to embarking on mini projects with linux - that just seemed a little too easy! :) |
Reboot, type dmesg and check that it says something like:
EXT3 FS on hda1, internal journal EXT3-fs: mounted filesystem with ordered data mode. If it does, then you're set. Incidentally, it's much easier to use ext3 because you CAN always mount it with any ext2 (not just ext3) utility (i.e. an old boot disk, that windows driver etc). But yes, it is MUCH HARDER to recover deleted files from journalled filesystems but not necessarily impossible - it depends on the design. Some journalled systems do stuff like zeroing deleted files etc. Recovering a deleted file is not something that you can just throw a utility like an e2undelete at, though. Incidentally, ext3 has three modes of journalling. One that's fast but not-so-safe, one that's completely safe but incredibly slow and one that's middling. The default is the middling (obviously). http://en.wikipedia.org/wiki/Ext3 I believe, though, that the worst that can happen with the default is that if you lose power while in the middle of writing a particular file, you will lose whatever parts of that file haven't made it to disk (as non-journalled systems do) but you shouldn't EVER lose the structure (i.e. you should never need to fsck, the linux equivalent of scandisk, such filesystems, should never get cross-linked chains and all that other rubbish and no other file should be affected). This means that, short of hardware failure, the filesystem shouldn't ever "break" or lose files/directories except for those that you were in the middle of writing (which may not be complete but should still exist). But then, if you were to crash halfway through writing a file on anything but a fully journalled, incredibly slow system, the chances that the resulting file will be useless (definitely if it's binary, almost certainly if it's text). So ext3 = readable as plain old ext2, but journalled enough for most purposes, secure enough that your filesystem won't corrupt if you do have a power outage, fast enough to still manageably use it, at the cost of a small bit of disk space. Reiser is better at storing lots of small files, generally quicker overall, takes less disk space, but you lose the easiness and convenience of ext3 (i.e. you can't convert an ext2 filesystem to a Reiser as easily as to ext3). |
Quote:
|
ledow, output of dmesg looks good:
EXT3-fs: mounted filesystem with ordered data mode. VFS: Mounted root (ext3 filesystem) readonly. EXT3 FS on hda6, internal journal cs-cam, Yes - i just thought it may have taken a couple of minutes at least to do its work. :) By the way the disk actually seems to be running better at a glance... I have not done any benchmarking or anything but in terms of 'noise' it seems quieter than it was before.... strange. |
Quote:
|
Quote:
|
All times are GMT -5. The time now is 01:15 PM. |