Converting ext2 to ext3 filesystem

[mikek147]

For those of you who would like to move to the ext3 journaled fs, the procedure is fairly simple. First the support needs to be built into the kernel. Although there are patches that can add ext3 support to other kernels, As of kernel 2.4.16, the support is built in. Recompile your kernel (2.4.16 and up) with "EXT3 journaling file system support(ENPERIMENTAL)" in "File Systems". Once compiled and installed, reboot.

Once rebooted, assuming all went well, log in as root. The command to add journaling to a specific partition, do:

tune2fs -j /dev/hdxy

where x = disk and y = partition. If your not sure what that is, just type the command 'mount' to see what is mounted.

Once you've been successful with this step, edit /etc/fstab and change then entry for the partition just changed from ext2 -> ext3.

Simple enough. Have fun. -mk

------------------
If it ain't broke,
Fix it till it is.

[rami3]

some questions mike - [*] - our linux file server serves about 1000 files a day and stores nearly 100000 different files with more being constantly added each day, would switching to ext3 benifit me and how?, is it an unnecessary risk?, what about the downtime when doing this?
i'm thinking here if it aint broke then dont fix it, but as part of my job is improvement of this network i would appreciate feedback.

Cheers RB

------------------
If builders built buildings the way programmers wrote programs, then the first woodpecker to come along would destroy civilisation.

[mikek147]

In a nutshell, a journaling fs is similar to log files in a database. If something happened to the database due to an incorrect shutdown, with the rollback/rollforward feature, you can make your db healthy again. Well the same thing happens upon reboot after a system crash. Instead of going through the time consuming ritual of running fsck, the system checks the journal, corrects just those items that may be wrong, shorting a reboot from several minutes+ to almost as fast as if nothing happened.

The larger the ext3 fs, the faster the reboot, in comparison to an fsck reboot. You have stated the large number of files you store. How much fs does it take up? Is it right for you, how long does it take you to recover from a crash? In the name of disaster recovery, this is something that should be known. What is acceptable?

The kernel rebuild is about the most time consuming part, including the reboot. The conversion from ext2 -> ext3 is done in the matter of minutes. The unmounting of an fs, editing /etc/fstab, and remounting the fs is more a matter of typing speed.

Is ext3 or any of the other journaling fs right for you? Only your can answer that. But I would suggest that, for your own sanity, read up on journaling before you make the big leap. It may be something you might want to test on a test machine first, so you can schedule the downtime more realisticly, if you should decide that this is something worth pursuing. Good luck. -mk



------------------
If it ain't broke,
Fix it till it is.

[rami3]

thanks for the info mike, from what you have said it seems that some of the users stations would benefit from ext3 more that the server, some work files are stored locally on the stations and as they are 'gettin on' a bit a fsck reboot can take up to 10 mins!!!!!!!!, our server never crashes (once in a year) and our backup/restore system on the server is very well covered, anyway i think i had better hit the books.................RB


------------------
If builders built buildings the way programmers wrote programs, then the first woodpecker to come along would destroy civilisation.