At the company where I’m working right now, I’m part of an architecture effort to come up with our standard design for RAC on Linux across the firm. There will be dozens or possibly hundreds of deployments globally using the design we settle on.
We’re internally debating whether or not we should include OCFS2 in this design right now, and I’m curious if anyone has arguments one way or the other to share. Our standard design on Solaris does utilize a cluster filesystem and we would welcome a similar design, but there are some concerns about the readiness, stability and future of OCFS2.
OCFS2 is being considered for these four use cases:
- database binaries (vs local files or NFS)
- diag top (11g) or admin tree (10g) (vs local files or NFS)
- archived logs
- backups
Other files will be stored in ASM.
I have seen mention in blogs such as http://bigdaveroberts.wordpress.com/ of something called ASMFS in 11gR2 and I’m wondering – will this feature (if included) have any impact on Oracle’s commitment to OCFS2 development? Could Oracle conceivably develop a whole new cluster filesystem and put their full weight behind it as they did for ASM storage, leaving OCFS2 as a lower priority for new features and improvements? Has Oracle demonstrated significant commitment to OCFS2 development and support in the past, and is this a mature enough technology for wide-scale deployment?
Just looking for opinions. :)
Because you have hundreds of systems to configure to standard, I would “avoid” anything that is either not solid or may change. And for that reason, I don’t think I would move forward with an OCFS2 solution.
I have not seen a lot of advances in the last few years. And we do see advances in other areas by Oracle, and other clustering filesystems from other vendors.
I love OCFS to “get the job done” with limited enterprise impact. Not sure if that is the direction you want to take.
LikeLike
You should also consider ZFS (fs coming from sun and available as a fuse module on linux) which could replace both OCFS and ASM if Oracle would decide to get and use the best technologies coming from their new acquisition (sun)…
LikeLike
ZFS is not a cluster filesystem – you can’t access it simultaneously from two hosts. It does have a lot of cool features though. I’m also a little geeked about btrfs – in fact it has many similar features to zfs.
In the end I recommended that the company use OCFS2, but primarily for business rather than technical reasons. They already had a standard design and extensive scripting for RAC on Solaris, utilizing a cluster filesystem. There is no cluster filesystem available for datafiles on Linux besides OCFS2. I felt that OCFS2 was a technically viable option and it would be similar architecturally to their Solaris environments – allowing them to leverage past Solaris scripting and design work to the maximum extent.
LikeLike
Ever considered ACFS since 11Gr2 ?
I like the idea for Oracle homes since it benefits the advantages from ASM.
jean
LikeLike
Thanks for stopping by, Jean-Francois! When I originally wrote this blog post, ACFS wasn’t released yet… and Oracle’s release timeline for ACFS didn’t work with our timeline for developing this architecture.
Anyway, I personally wouldn’t recommend ACFS yet for a major architecture effort like this. Definitely good to start trying it out in a few places — but I’d definitely wait a little while before doing any wide-scale deployments. OCFS2 has been around for a bit longer and I do think that it’s a viable option for wide deployment.
But for what it’s worth, some other engineers disagreed with me about OCFS2 and the project eventually ended up excluding any linux cluster filesystem. They just re-engineered the design to use local storage only (and ASM for data). Personally I was voting for OCFS2 but the final design wasn’t so bad.
LikeLike