If your apps do not care about atime, then noatime is helpful. data=writeback should performs better than data=ordered. But there is a small chance that files having trailing nulls if a node were to reboot after a journal commit but before a data flush. This is documented in the manpages and the user's guide. I don't believe nodiratime is a standard mount op. May be specific to a fs. noatime should be enough.
So I would use noatime. Journal mode will depend on the app/use case. On 05/24/2011 11:42 PM, Stefan Priebe - Profihost AG wrote: > @Sunil > So could you recommand generally mount options to set? At the moment > i've set none. > > So is it always a good idea (don't need atime) to set > data=writeback,noatime,nodiratime any other options? > > Stefan > >> Did you set the mount option on both nodes or only on the node >> on which you were doing the ls? >> >> Setting it on both nodes, or on the node that is doing the cp should >> solve the perf issue. What's happening is that the ls on node2 is >> forcing >> node1 to journal commit. With the ordered data journal mode, the data >> is flushed on commit. Switching to writeback will allow it to commit >> without flushing the data. _______________________________________________ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
