ext4 appears to me as a very promising filesystem and a welcome successor to the somewhat aged but still venerable ext3.
i've read through the nice introduction at http://groups.google.com/group/linuxkernelnewbies/browse_thread/thread/9798c3fd512fcbb1/ad9887da58d68c11?hl=en and became even more curious. while trying the ubuntu 9.04beta 64bit which comes with a kernel based on 2.6.28.8 i decided to give ext4 a try and use it for a data- partition that gets mounted on-demand somewhere below /home/me/ . i needed to manually format the partition and used mkfs.ext4 to do so. during this step, i remembered some parameters i used once for a similar task but with ext3 as a filesystem like the inode-presets filesystem-type (mkfs-parameter '-T') and blocksize (mkfs-parameter '- b'). for my old ext3 data-storage-partition i used a blocksize of 4096 which is implied by the usage of specifying a filesystem-type of 'largefile' or 'largefile4'. files stored on that exisiting partition (and will be more or less the same with the new ext4 partition) have an average size of around 6MB so at least for ext3 a filesystem-type of largefile4 seemed favourable. but i am wondering if this applies to ext4 as well?!? i tried 'mkfs.ext4 -m 0.02 -T largefile4 -v /dev/ targetpartition' (since its a storage-partition a reserved blocks- percentage of 0.02 should be ok i guess) and found that this results in a blocksize (and a fragmentsize) of 4096 which leads to a total of 236149411 blocks 230624 inodes 7207 block-groups 32768 blocks and fragments per group 32 inodes per group for my 900GB partition. since i use a 64bit version on a 64bit system i should be able to use larger blocks (have to check the pagesize again...), ext4 allows blocks to be up to 64kb instead of only 4kb. also when running mkfs.ext4 without any filesystem-type specified a blocksize of 4096 is chosen, so i wonder if i should manually override this to lets say 16384 or even 32768 and lower the number of inodes accordingly (/4 or /8)? another question is if keeping the number of inodes rather small etc makes still sense with ext4, but i don't see a reason why this shouldn't be the case. one thing i don't have a clue of is how to determine a reasonable groupsize. anybody got an idea? suggestions are warmly welcome ;) thanks a lot, shnaxe
