On 6/24/2011 5:31 PM, John Drescher wrote: >> I would use 'xfs'. I believe samba was originally developed >> over xfs, so it's likely the ea-suppot and acl support has had the most >> testing there. Especially if your file server is setup with a UPS, then I'd >> strongly recommend it. If not, ext4 might be safer (with write >> through). It will be slower, but safer. >> >> With a UPS, XFS's default 'write-back', will give the fastest >> performance for large file writes (I think reads as well). It's worst >> performance is on "removing" large numbers of files, as that is pretty >> much a synchronous operation... > > I would just use ext4, it does not have the ext3 large file slowness > or xfs slowdown with lots of small files.
"xfs slowdown with lots of small files" is no longer true. To be accurate, the complaint was never with "lots small files" but with metadata write performance, i.e. deleting, renaming, changing attributes, etc, of lots of any sized files--operations that saturate the log journal. This poor reputation was gained long ago because XFS yielded relatively poor performance with operations such as "rm -rf" on a kernel source tree. Such an operation is metadata write intensive and previously would bring the XFS log journal to its knees, saturating the physical IO channel(s) to the disk subsystem, creating a severe bottleneck. Today this type of operation is as fast as EXT4 thanks to Dave Chinner's ingenious delayed logging patch. It in essence pushes much of the previous journal IO operations into memory, consolidates the log writes, and thus decreases actual disk IO eliminating the bottleneck: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=Documentation/filesystems/xfs-delayed-logging-design.txt -- Stan -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/options/samba