[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: 2GB of Waste? How can it be? -- sysadmin cutting in

[ SysAdmin, non-fs guru cutting in ]

"IT3 Stuart Blake Tener, USNR-R" wrote:
> In the instant case, I am creating a filesystem for /home
> directories, thus, a 5% reservation for root is not needed here, so I
> shall take your advice and return it to the general free space pool.

Ummm, the 5% reservation is to prevent the high levels of
fragmentation that occur when the filesystem is near full (something
that I wish Windows would adopt as standard too ;-).  It is also to
keep your system from "hanging" if system/root processes need to
write to the filesystem, so they aren't at the "mercy" of users
filling it up.  And there are a few other, very important reasons

So, I highly recommend you do _not_ remove it.  It's saved my @$$
several times when my users didn't bother to care.  ;-P

> Why cannot ext2/ext3 allocate inode space on a dynamic basis?

Before you ask that, ask yourself:
"Do I want a proven filesystem to be mucked with?  Wouldn't it be
easier to check out ReiserFS, XFS or JFS instead?"

I, for one, don't want Ext2/3 changed because I like its _proven_
reliability _and_ (more importantly) recoverability.  I'm sure
Tweedie & co. will get to more radical Ext2/3 changes (Ext4?), but
I'll take reliability and recoverability any day over features.

On that note, SGI's XFS filesystem shares much in common with Ext2/3
features (e.g., POSIX ACLs) and realibility (proven on Irix), in
addition to on-the-fly inode creation, etc...  If you want then, use
XFS.  SGI's releases for RedHat, based on the same RedHat RPMs,
allow you to use both Ext3 and XFS (as well as Ext2).

> More importantly why is it not possible to amend the number of
> inodes an ext2/ext3 filesystem has after the point of doing an mke2fs?

Again, ask yourself the above question.  ;-P

> One other question, as well:
> I modified my RedHat installation to boot up using a root
> filesystem of ReiserFS, but when I try to use devfs=mount along with an
> initrd, it seems to give me some errors (initrd= something or other
> error, I will write it down next time). Any ideas?

Bit of advise for you that works well for me:  
Always make / Ext2.  You should be separating out at least /tmp,
/var, /usr and /home anyway, so / doesn't need to be that big. 
Hence, journaling is not a big deal.  If / is Ext2, I can boot any
disk and read /.

Regarding ResierFS, such questions are best for the ResierFS lists. 
There are some issues with using LILO and ResierFS as /, but GRUB
has a ReiserFS driver to overcome them.

> What command do I give to tune2fs to change the space
> reservation percentage?

_Always_ consult the oracle of man before you post ... ;-P

 $ man 8 tune2fs
 -m reserved-blocks-percentage
          Set the percentage of reserved filesystem blocks.
 -r reserved-blocks-count
              Set the number of reserved filesystem blocks.

> Also, I noticed that if I turn on ext3 journaling, I can no
> longer resize the partition with partition magic. So I am curious, how I
> can completely revert an ext3 partition back to ext2 (temporarily) so I
> can resize it, and then make it ext3 again. Is there a command line
> parameter for tune2fs which will reverse a "tune2fs -j"?

I believe both the FAQ and archives have answers on this.

BTW, I wouldn't trust resizing Ext2 with Partition Magic --
especially when used on newer kernels.  I've seen a number of people
getting filesystems that are barely recoverable with fsck

-- 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.

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]