I have just started using aufs. I'm running it on a fresh, minimal install of Debian stable. I'm mount 3 ext4 formatted drives as rw with the following line in my fstab:
aufs /mnt/all aufs nodev,create=mfs:600,br:/mnt/.hidden/2Ka=rw:/mnt/.hidden/2Kb=rw:/mnt/.hidden /taft=rw 0 0 $ uname -a Linux rawls 2.6.32-5-amd64 #1 SMP Wed May 18 23:13:22 UTC 2011 x86_64 GNU/Linux I'm not running much in the way of software at this point. I sshfs in to the machine to access the files locally and am running rtorrent. I noticed the data corruption with rtorrent. I've replicated the corruption several times now using these steps: 1) Download new data in rtorrent, write directly to the aufs. 2) Shutdown rtorrent 3) Reboot system 4) That previously downloaded data no longer passes a hash check. The data is NOT corrupted before the shutdown, only afterwards. After the failed hashcheck, if I redownload the data and reboot again, it is re-corrupted. Furthermore, I noticed a significant amount of swapping (on a machine with 8 gigs of RAM and not much running on it). I don't know if this might have something to do with rtorrent's use of mmap and aufs. Is there a known problem with using rtorrent and/or mmap and aufs? Is this an unsafe way to be using aufs? Note that not all writes are being corrupted. I rsynced about 300 GB of data to the 3 drives, and that data appears to have maintained its integrity. I'd really like to be able to use aufs to union these file-systems together. I really appreciate the round-robin writes. -- james
------------------------------------------------------------------------------ EditLive Enterprise is the world's most technically advanced content authoring tool. Experience the power of Track Changes, Inline Image Editing and ensure content is compliant with Accessibility Checking. http://p.sf.net/sfu/ephox-dev2dev