In short, no. Unfortunately, I'm really busy with my very interesting (all-)day job and have 2 small children and a wife that take up any remaining hours I have. I still reply to email regarding ext2resize, and if anyone wants to work on it I'm happy to help them along. There is some hope that Stephen Tweedie will finish off the last remaining bits to get a new release out which supports ext3 (see below).
Very little, actually. The core ext3 kernel patch is already done for 2.4.x and 2.6.x kernels, but no up-to-date patch is available. The ext2online support for working with such a kernel is also finished. The only remaining bit is the ext2prepare support for allowing existing filesystems to be grown beyond the next multiple-of-16GB boundary. As a hack you can patch mke2fs and it will prepare new filesystems for growth up to a factor of 1024x of the original size.
Yes. Or at least it's supposed to be :-) But if you worry about your data then you should have a backup. You should have a backup even if you are not resizing - what if you accidentally delete files, or your hard disk dies (along with power supply, hard disks fail the most).
Pretty much. I've had several reports from users who resized their 1-20GB filesystems successfully. I've also had a report from somebody who successfully resized a news spool from 72GB to 144GB!! But anyway, if you use this program, be sure to make backups. Or make sure you don't have very important data on the to-be-resized filesystem. If you decide to test it, please send me a success/failure report (include the version number and start-end sizes please).
No, not at all. Just use ext2resize on a loopback file and it will work fine. See ext2resize man page for an example of how to test with a loopback file.
If I did it right the fs will be more or less recoverable if it didn't crash while moving the inode tables (this operation is pretty non-atomic). I could work around this but I'm not doing this yet. Note that I say "more or less": if you're worried that a power failure will thrash your partition you probably have vital data there, and you should make a backup!
Yes you can, using the ext2online program (part of the ext2resize package). However, you need to have a kernel patch in order to do this. Please read the documentation in the package.
Yes you can, but only using the CVS version of ext2resize or ext2online. You still need to have a kernel patch in order to do this. Please read the documentation in the package.
Is it slow? Well, first you might want to look at tune.c for tunable parameters. Try to increase the size of the relocator pool. Not too big; it should still fit in core. This should make it faster. If it is still slow, please compile ext2resize with profiling turned on (-pg) and send me the output of gprof.
Because it uses all kinds of heuristics, shortcuts and optimizations. View the source, Luke!
There is a generic filesystem resizing package, parted, which is working towards this goal. It uses the ext2resize code to handle ext2 filesystems, and can also do at least VFAT, and probably others.
No, no plans for this. Stephen Tweedie's defrag tool only works on 1kB block filesystems, so it is generally not useful anymore. Andrew Morton has written a patch to provide the hooks for a defragger which can run while the filesystem is mounted, but it needs a user-space tool.
Probably not worth the effort.
It might be possible to go from a 4k blocksize to a 1k blocksize, but it would be very difficult to go the other way. Unfortunately, it is most useful to go from 1k to 4k blocksize. It would be very very hard, and would probably require that you make a backup first, at which point you may as well just restore the backup into a filesystem with a larger block size.
This is a bit easier, but very sensitive to crashes. I have a scheme for this, but don't hold your breath, I'm very busy concentrating on other things.
Yes, it can. But if you're interested in this, make sure you visit the Logical Volume Manager for Linux page. 'Tis really neat stuff.
Now, any good web browser can be configured to display the page in the way the *user* wants - if you want the page to look green, configure your browser so it displays the page in green. A HTML page as such has no appearance, although many web designers try to force things upon you to try and make their pages look cool.
Patches, publicity, supportive emails, whatever. Just let me know you're there. I like the idea that there are actually people using this program now.
As of 991005 I've had about 300000 downloads total.
Not much - it makes me wonder if the code is totally bug free, or if nobody is using it.