Hi,

>> I have enough space to do this with my current provider (Omnis), but
>> they have some weird limitation about total number of files, which I
>> keep bumping up against

> Not a solution, but hopefully just an explanation. They're probably  
> trying to
> stave off inode limitations of their chosen filesystem. Filesystems  
> have a
> maximum number of files placed upon them at formatting time, and  
> it's a
> nightmare if they run out of inodes.
>
> I run a proxy for informationfreeway, and it used to do a Gig a day of
> traffic. The cache would fill up with millions of files, and could  
> easily eat
> up all the available inodes within a couple of weeks. Free space  
> was still in
> the tens of Gigabytes range, but no new files could be created.

If you can live with a performance trade-off, you can always create a  
huge empty file, make a file system on it, and loopback-mount that,  
like so

dd if=/dev/null of=my-large-file bs=1024000 count=1000

gives you a file of about 1 gig, then

mkfs.ext3 ./my-large-file

("my-large-file is not a block device, proceed anyway..." answer yes)

then

mkdir my-new-dir
mount -o loop ./my-large-file my-new-dir  (needs root privs)

There you are, a new file system under my-new-dir. And if you specify  
"-T news" when making your ext3 filesystem (or use a Reiser fs right  
away), you have more than enough capacity for little files.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail [EMAIL PROTECTED]  ##  N49°00.09' E008°23.33'



_______________________________________________
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk

Reply via email to