[leafnode-list] Re: Run out of inodes

[Matthias Andree]

Am 16.05.2009, 13:08 Uhr, schrieb Enrico Wiki  
<enrico.c.ml.address at gmail.com>:

> On Wed, May 13, 2009 at 10:50 PM, Matthias Andree  
> <matthias.andree at gmx.de>wrote:
>
>> Am 13.05.2009, 19:54 Uhr, schrieb Enrico Wiki <
>> enrico.c.ml.address at gmail.com>:
>> Well, one article, one inode. Plus around 1,100 overhead for  
>> directories.
>> You may want to create a separate file system with higher inode density  
>> for
>> the spool if you're so tight on inodes. mke2fs.conf suggests one inode  
>> per
>> 4,096 bytes for news spools, and the default is 1/4th of that (1 inode  
>> per
>> 16 kibibytes).
>
>
> Thanks. For now I am trying out reiserfs, which allocates inodes
> dynamically, so you don't need to set a number when you create the f.s.

Reiserfs v3 is IMHO obsolete. Tail packing can be slow (ovvercome that  
with the "notail" mount option, at the expense of some disk space) and in  
my perception it is not as solid as ext2 or ext3 are. There are reports  
that its journalling style is inferior to full-block journalling as done  
in ext*. Proceed at your own risk.

I'd suggest ext3 for Linux, UFS2 for FreeBSD, logging ufs on Solaris. For  
Linux, running mke2fs with "-T news" on a partition will provide more  
inodes (currently 4x of what you'd get with default parameters).

>> Is Ubuntu doing strange aliases, libc or kernel hacks, or have you used  
>> a graphical file manager? For "rm" should not move files to trash...
>
> just graphical file manager

Ah, ok. Emptying the trash, or for some desktops, pressing Shift+Del  
deletes everything rather than moving it into the trash.

HTH

-- 
Matthias Andree