Zbr's days.

About :: TODO :: Blog :: RSS :: Old blog :: Projects :: GIT :: Gallery :: Notes

Fri, 15 Jun 2007

BTRFS by Chris Mason at Oracle.


There are number of features currently not implemented, and which limits design:

  • Absence of delayed allocation.
  • This means that fragmentation will kill a filesystem, even if it is extent based (which is just a block size increase). Currently btrfs allocates new on-disk chunk for each 8k write.
  • Bad tree locking
  • Which does not allow multithreaded writing.
  • Some minor nits like absence of sync, direct and mapped processing, which is unlikely to be design problem though.
Actually, I plan to watch this project, but currently limit my filesystems interests to the block layer - I still want to implement distributed storage with raid-like (not exactly raid, since for higher order checksums (like 32 bit crc) it requires slow Galois-field multiplications, but something like that) functionality, which could be put under the filesystem and allow to put usual filesystems on top of highly-sized arrays distributed over the net.

/devel/fs :: Link / Comments ()