Highly available filesytems

Thursday, 24. 11. 2005  –  Category: all, sw

A hosting project I’m working on needs a large filesystem to be available on multiple hosts and to be continually mirrored.

  • A pair of NFS servers rsyncing master to slave – feeble
  • Coda – looks good but project appears dead (though has current kernel support on both FreeBSD and Linux), 32bit limitation on metadata restricts fs size.
  • Lustre – shiny, very current, Linux-only

I will play with Lustre on a pair of lab boxes

Really, I want to do something with ZFS – it has features galore.

2 Responses to “Highly available filesytems”

  1. Platypus Says:

    Yes, ZFS has features galore, but it’s not a cluster filesystem. Sun has one, but I imagine being locked into Solaris would be even worse than being locked into Linux. If being Linux-only is not a problem, you’d probably also want to consider GFS – originally by UMN and Sistina, since bought out by Red Hat.

    If you’re considering Coda, why not consider OpenAFS instead? The main difference between AFS and Coda was disconnected operation, which would seem to be of little interest based on what you’ve described. OpenAFS is the lineal descendant of AFS, is supported on many platforms, etc.

  2. lemon Says:

    ZFS not a cluster fs, ack – just something it’d like to check out for myself i guess. Cheers for the other pointers, I’ll check them out.

Leave a Reply