"IT3 Stuart Blake Tener, USNR-R" wrote: > The only reason I am not using EXT3 (I think using the more > advanced journaling methods, it is a stronger product), is that I am > very concerned about maximizing the space I get out of my laptop hard > drive, that is my main concern, next to the filesystem being stable (I > consider EXT3 and ReiserFS to both be stable at this point). > Perhaps it is time for me to evaluate again EXT3, and look at > how much the space difference is. It sure is convenient to use the > filesystem that the distribution wants you to pick as the default, and > EXT2/EXT3 is well thought of. Er, free disk space is not going to be a fs-implementation issue. ReiserFS has a B*Tree that more than make up for its "inode efficiency" issue, although it depends on what is stored. I think the B*Tree is pre-allocated/reserved in V3 (anyone?), which eats up several tens of MB anyway. But this adds up to pennies in the end. But even if you still consider it "essential," then XFS is a _even_better_alternative_ to ReiserFS V3! In addition to its dynamic inode allocation, XFS can store a small file's data in its inode, for space savings (i.e. doesn't use a data block). ReiserFS still requires you to eat a 4KB (or whatever the basic size is) data block, _in_addition_ to eating _both_ inode pointer space _and_ space in the B*Tree. XFS also stores symlinks in the inode as well (saving a datablock versus Ext2/3), although ReiserFS stores symlinks in its B*Tree for about the same "cost" in the end. XFS does other inode/freespace management much better than ReiserFS V3 IMHO (except for lots of deletes, then XFS does slow). [ Information according to a Linux Gazette #55 comparison of the four major JFS' for Linux, located here: http://www.linuxdoc.org/LDP/LG/issue55/florido.html ] And as you found out, the "reservation" of disk space can be changed on-the-fly with "tune2fs", although don't expect good performance out of _any_ filesystem that is less than 5% free (like _detrimental_ ;-). > It is also that I know ReiserFS is coming out with V4, which is > supposed to have encryption built it as well, and again, that > will also be important to me. ReiserFS V4 is _far_ from being adopted in the stock kernel AFAIK. Heck, one of the reasons why ReiserFS V3 had "trouble" getting into the stock kernel, from what I understand, is that Linus required Namesys to "tie down" the actual structure. That's not to say ReiserFS V4 won't be excellent, and I like a lot of what I see in it (and ReiserFS in general). But from your "viewpoint" that "RedHat should put it in the installer as standard" -- sorry, not going to happen, not in-line with what most RedHat consumers (at least AFAIC) want. Hence which makes your posts here somewhat ... how should I say ... less-than-interesting??? -- Bryan P.S. Anyone should feel free to correct me where my information/assumptions are wrong. -- Bryan J. Smith, Engineer mailto:b.j.smith@ieee.org AbsoluteValue Systems, Inc. http://www.linux-wlan.org SmithConcepts, Inc. http://www.SmithConcepts.com --------------------------------------------------------- 1999 IRS Data: The top 1% of income earners pay over 36% of the taxes, but have less than 20% of the total income.
