My project wish-list for the next 12 months

[Peter Ross]

Scott Long, Wed Dec 1 14:02:28 PST 2004

> 5.  Clustered FS support.  SANs are all the rage these days, and
> clustered filesystems that allow data to be distributed across many
> storage enpoints and accessed concurrently through the SAN are very
> powerful.  RedHat recently bought Sistina and re-opened the GFS source
> code, so exploring this would be very interesting.

As we see there are some unfinished pieces (BTW: The freebsd-cluster list
is very quiet).

I browsed through some of them. It is difficult to judge how far they away
from beeing mature.

There is:

1) 2x AFS
   (BSD project, OpenAFS)

2) iSCSI
      The Lucent code

      Peter Blok seems to have some ideas comparable to mine when I started
      to play with iSCSI at FreeBSD when I was unemployed last year.
      (Unfortunatelly not long enough;-)

      My naive idea for network mirroring to get basic SAN multipath
      functionality is to combine vinum + iSCSI. Is it achievable?

3) NFSv4

      NFS Version 4 is part of FreeBSD now. There is an idea to implement
      replication but AFAIK it is not implemented yet (I have to check).

      At least there is a whitepaper:
      http://www.citi.umich.edu/projects/nfsv4/reports/replication.pdf

      It could be used as a frontend for concurrent access (NFSv4 provides
      _real_ locking including byterange locking etc.)

There are some ideas to implement concurrent access in UFS.. Is it the
right place? (Hint: There are many Linux filesystems with beautiful
features but it is hard to find a reliable one - possibly because the
featuritis makes them too complex?)

Maybe it is better to have one specialised filesystem (see Veritas as
already mentioned). Or a layer above the "real" FS.

BTW: AFAIK ReiserFS (suggested as a FreeBSD port) does not support ACLs..

Regards
Peter