Hi,
At first, i hope that you could excuse cause my english isn't better -
but i'm brazilian, and here we talk in portuguese.
I am a fault tolerance graduate student, and my work deals with
development of fault injection techniches and tools for aspects related
with Linux operating systems. Recently one of my dear colleagues, Mr
F'abio Olive (by Conectiva Linux distro), presented his works in the 8th
International Linux-Kongress, on Netherlands - talking about one fault
injection tool for communications systems, using Netfilter. And i, by
the way, are interested on XFS because my research sub-topic are the
journaling filesystems
;-)
Please, could you tell me what aspects you think that should be more
interesting for a fault injection tool? I think that XFS is the best of
the journaling filesystems that exists - rather than ext3, jfs. So, in
despite of my opinion, i think that it could (and should) be proved
under an academic point-of-view.
I wanna know what you think that should be tested more carefully? What
part you think that has more problems (solved and pending)??
I've seen mailing list archive, describing problems (which are solved
very quickly, congratulations by that), but know i 'm accepting
suggestions. My scenario is considering a RedHat 7.2 installed using the
SGI cd-rom installer, and recompiled with XFS 1.1 to enable the
CONFIG_XFS_DEBUG on kernel. I've been installed the latest version of
tools (xfsdump, xfs_repair, etc)available on the FTP website. I don't
made any synchronization with CVS.
Finally, my congratulations for all by the well done work. I think that
SGI made a great contribution to open-source world wide community
licensing the XFS filesystem under GPL.
;-)
Cheers,
Leonardo Mello
|