> There's thing that have always been wrong. Always from whose perspective? Everything you list is a result of a historical compromise from a time where there wasn't any other way to accomplish the task, or is an artifact of dealing with closed source components, and refactoring for refactoring's sake is just a waste of resources. We're dealing with 25-30 year old code here; standards were different then, and design by expedience is a hard thing to recover from.
Regardless, AFS is still reliable, supportable and viable, and people use it in general purpose environments more often than OCFS. I expect that may change over time, but that's what I see out in the real world at the moment. OCFS is just too new to have developed a large following yet. > > It would be nice if it tracked the openafs tree a little closer, though. > > The in-kernel code has needed a lot of work so far to avoid causing > > miscellaneous problems in large deployments. > > I don't think there is any point in tracking OpenAFS code. It's of > absymal code quality, and under a strange license that wouldn't even > allow to reuse a piece of good code if you managed to find it. On the other hand, since the OpenAFS code is where fixes and functional enhancements are being done for AFS, it's a bit on the rude side to not track it, especially when it would allow one to avoid being a support PITA. Interoperability is a wee bit more important in non-trivial enterprise architecture, and many of the problems with the in-kernel code I mentioned have more to do with not being aware of the whole picture than coding errors. The Arla folks seem to manage not to be a PITA; it seems more practical (and productive) to me to work together rather than in isolation. The IBM license for OpenAFS is what it is. Is it perfect? No, but it's what IBM would agree to, it's the one we have, and it's better than a totally closed source approach. It's no worse than the BSD license, which is (considered objectively) not all that awful. As to code quality, based on close to two decades or so experience with the CMU and then the Transarc AFS code, I can say that the OpenAFS code is certainly a dramatic improvement from the Transarc code that IBM tossed over the wall, and has been steadily improving over time. You may not like it yet, but it's a lot better than it used to be -- and isn't incremental improvement the goal of open source? If you've got time to spare, assistance in perfecting the code to your standards is always welcome. We've drifted away from Linux-390 and cluster filesystems, though. I'd be happy to continue the discussion offlist. -- db ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390