Dr. Tweedie, et al.: I recently formatted a partition using EXT3, and after a "df -h" I get 14GB of space. When I reformatted the partition with ReiserFS, and did a "df -h" I got 16GB of space! Now the partition was setup to be 16GB via fdisk, so 16GB is correct. However, why does EXT3 loose 2GB of space? The journals cannot be that big!?! Very Respectfully, Stuart Blake Tener, IT3 (E-4), USNR-R, N3GWG Beverly Hills, California VTU 1904G (Volunteer Training Unit) stuart@bh90210.net west coast: (310)-358-0202 P.O. Box 16043, Beverly Hills, CA 90209-2043 east coast: (215)-338-6005 P.O. Box 45859, Philadelphia, PA 19149-5859 Telecopier: (419)-715-6073 fax to email gateway via www.efax.com (it's free!) JOIN THE US NAVY RESERVE, SERVE YOUR COUNTRY, AND BENEFIT FROM IT ALL. Monday, February 04, 2002 10:02 AM
Hi, On Mon, Feb 04, 2002 at 10:05:48AM -0800, IT3 Stuart Blake Tener, USNR-R wrote:> Dr. Tweedie, et al.: > > I recently formatted a partition using EXT3, and after a "df -h" > I get 14GB of space. When I reformatted the partition with ReiserFS, and > did a "df -h" I got 16GB of space!ext3 uses exactly the same rules as ext2 here. By default, 5% of the filesystem is "reserved" for root --- tune2fs can modify that. There is also the fact that ext2 and ext3 preallocate space for inodes: if you expect only to have a few large files, "mke2fs -i" can be used to reduce the amount of inode space preallocated.> Now the partition was setup to be 16GB via fdisk, so 16GB is > correct. However, why does EXT3 loose 2GB of space?It doesn't --- the space is just reserved, either for root access or for the inode data structures. reiserfs also needs space for data structures analagous to inodes, but it allocates that space dynamically so you don't see it on a newly-created filesystem.> The journals cannot > be that big!?!No, they should only be a few tens of MB at most. Cheers, Stephen
"Stephen C. Tweedie" wrote:> ext3's primary goal is stability. There are plenty of alternatives if > you want bleeding edge.As I mentioned before, SGI makes a nice .iso image you can download and install with RedHat 7.2. They have a modified Anaconda installer, and have even fixed a few bugs for RedHat here and there.> Actually, no, it isn't --- you have flat out insulted the developers > involved. What on earth do you imagine we'd have to gain by promoting > one open-source module over another for such petty reasons?RedHat, despite having Ext3 kernels in-house, kept it out of its release for the longest time -- a good 18 months or so. Same deal with their kernels, they lag other distros for months _because_ of testing. RedHat gave their "fair shake" of ReiserFS, just like I did over two years ago. No dice AFAIC. I've played with it several times since, but don't have the time to deal with different patches for different services and support.> Red Hat tries to offer a reliable and stable base OS, not to add every > possible bell and whistle to the distribution, and frankly our users > get better value out of our development time if we spend it > debugging the supported features than adding code for stuff we don't > support. The big advantage of open source is that you don't HAVE to > wait for us to add such features.I think that last statement is the key to RedHat. Especially when you add in the fact that almost everything they produce is 100% GPL. Cannot say the same about a number of other, major, and even well-respected distros. RedHat has stated why they don't support ReiserFS, and it is extremely logical. Accept it and find another vendor who does if you don't like it. Or, again, are you missing the whole point why you, I and many others stick with RedHat? -- Bryan -- 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.
Stelian Pop wrote:> Completly OT,No, I started it, my bad. ;-P> but do you also know about http://apt-rpm.tuxfamily.org ?I figured any Apt repository for RedHat would probably need to be independent. Thanx for the link. Contact me directly with any further correspondence. -- Bryan -- 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.
"Stephen C. Tweedie" wrote:> Come on, reiserfs didn't even have a working fsck when SuSE first > shipped it.Not to mention that while SuSE was using ReiserFS internally on a NFS network, they had to use user-space NFS (or a heavily patched, non-knfsd-based kernel version). -- Bryan -- 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.
On Tue, Feb 05, 2002 at 01:26:29AM -0000, matt dickinson wrote:> <snip> > > That requires changing very sensitive data, and could easily > > result in data > > corruption. You'd also increase fragmentation. > </snip> > > Is it possible to defrag(ment)?Utilities exist for such a purpose (look on http://freshmeat.net), but little purpose exists for them. Unix filesystems by their very nature resist fragmentation very strongly. It is not uncommon for even a very full, very heavily used filesystem to keep its fragmentation under 15%. Also, because of advanced read-ahead, disk buffers, and file caches, Unix also tends to be much less affected by fragmentation than, say, M$ Windoze of any variety. -- -- Skylar Thompson (skylar@attglobal.net)