Author: Frank Naumann (wh58-508.st.uni-magdeburg.de)
Date: 08-07-2001 12:52
Hello!
>With very small files yes. For normal use >(the two lower tables) the difference
>is not so big. As I mentioned in my oriignal
>post.
Hehe, you don't mentioned what type of use. Maybe he want to store thousands of small files (this can easily happen on development, html projects and so on).
But at all the ressource utilization of ext2 is much better (automatically load balancing, block preallocation, dynamic names in directory allocation and so on).
There is still the 65536 inodes limit too.
>From MinixFS documentation:
>6. It is possible to make *huge* partitions, >theoretically the limit is 4096Gb
AFAIK this was never really tested. From my own experience during the rework on the MinixFS I can't say how reliable such big MinixFS partitions are. The sources are not very clean in this case and mix signed/unsigned short/long data types in lot of places.
>Now, my 2.7g MinixFS partion is impossible >regarding you. And well, it's fun to do the >impossible ;-)
It may work. But I'm absolutly not sure how the minix fsck perform on such partitions. This wasn't tested too and the checker was written lot of years ago. You have ever filled up your MinixFS partition to the limit?
At least the ext2 tools for checking/repairing are much better than the MinixFS tool.
Frank
|