Results 11 to 15 of 15
Mind you, it has to be mounted read-only. Otherwise it would be impossible to fsck the root filesystem....
- 07-03-2005 #11Linux Guru
- Join Date
- Oct 2001
- Location
- Täby, Sweden
- Posts
- 7,578
Mind you, it has to be mounted read-only. Otherwise it would be impossible to fsck the root filesystem.
- 07-03-2005 #12
Thats what i though i saw about fsck - and now budman7 has told me where i saw it - probably tried it a while back and saw that message. Funny something that importanmt isn't in the man page for fsck!?
I wonder if you could say why? What doies it offer? My reading stack is overflowing right now - don't mean to be lazy.
Originally Posted by Dolda2000 - Clouds don't crash - Bertrand Meyer
registered Linux user 393557
finally - hw to brag about - but next year it will look pitifully quaint:
Athlon64 X2 3800 - 1G PC3200 - 250G SATA - ati radeon x300
circa 2006
- 07-03-2005 #13Well ext3 has journalling. This baisically means that the filesystem journals the information on it, so if there is a unclean umount etc, it is easier to get the system back on track. Some info on it.
Originally Posted by twoHats
- 07-03-2005 #14Linux Guru
- Join Date
- Oct 2001
- Location
- Täby, Sweden
- Posts
- 7,578
Sorry, I just assumed that you knew. As dylunio says, ext3 has a journal.
Originally Posted by twoHats
Technically, that means that every hard drive transaction that must be commited atomically is first stored in the journal, and then commited to the actual filesystem. If the system goes down uncleanly or the filesystem is otherwise uncleanly unmounted, the transactions can still be carried out atomically using the data in the journal.
As dylunio says, it prevents corruption due to unclean unmounting. Since the transactions are replayed from the journal, no fsck'ing or anything is needed. There are several journalled filesystems, including but not limited to ext3, Microsoft's NTFS, XFS, ReiserFS and JFS. Linux's ext3 is widely heralded as having the best journalling. It lacks a small bit in some performance benchmarks, and ReiserFS is better than almost anything when it comes to fragmentation, but you will almost never get a corrupted ext3 filesystem.
Otherwise ext2 and ext3 are identical. The only difference is the journal. Since they're so similar, it is very easy to convert an existing ext2 filesystem to ext3. Just run "tune2fs -j /dev/thepartition", while unmounted, and then mount it as ext3. Also, remember to change ext2 to ext3 in your fstab.
- 07-04-2005 #15
Thanks again ...
Thanks again for a great explanation - I used to work with a product called tuxedo for UNIX and another proprietary Startus product to do similar things with dbs - just didn't know that term (Journaling) as my tech has cobwebs
. I assume there is a small price to pay in speed?
I will gladly look things up from here - Thanks for the tip and all the help.
btw - I figured when i did the unclean unmount - when i created the partition, i wasn't thinking - whoops - Knoppix knows nothing of my new partition - I am so used to it recognizing all partitions and unmounting them - gah, what a move
I'll see you on the forums -- Clouds don't crash - Bertrand Meyer
registered Linux user 393557
finally - hw to brag about - but next year it will look pitifully quaint:
Athlon64 X2 3800 - 1G PC3200 - 250G SATA - ati radeon x300
circa 2006


Reply With Quote
