OK, I was missing the real nature of the problem; As it turns out, it wasn't the Server OS reporting an error, but rather just the Windows SMB/CIFS client throwing a fit. To wit, I hadn't actually tried writing files locally; Instead, my family had just been sending their various backups to the server, and Windows was telling them that the server was full, and I believed it... Then I read https://btrfs.wiki.kernel.org/index.php/FAQ#Why_does_df_show_incorrect_free_space_for_my_RAID_volume.3F which states that it's actually very difficult to tell how much space is available, so if Windows asked, SAMBA wouldn't know how to answer...
Interestingly enough, sending the files from a Linux SAMBA client presented no issue. SO, this is PROBABLY a SAMBA Server on BTRFS bug... But it only presents with Windows clients. At the moment, I'm just telling the family to send files in smaller chunks, or to use SCP; Both seems to sidestep the problem, but isn't a REAL solution... Any ideas how to deal with THAT problem? On Wed, Dec 17, 2014 at 9:58 PM, Tyrell Jentink <[email protected]> wrote: > > I'm using BTRFS on a Fedora 20 server. My kernel version is 3.17, and my > BTRFS version is v3.17. My pool is composed of 2 identical 3TB drives; Some > details on my system: > > #btrfs fi show >> Label: none uuid: 4f4b4c32-dc58-4288-a0d1-98db185dd429 >> Total devices 2 FS bytes used 101.25GiB >> devid 1 size 2.73TiB used 103.03GiB path /dev/sdc >> devid 2 size 2.73TiB used 103.01GiB path /dev/sdd >> Btrfs v3.17 > > > Which is consistent; I have the two 3TB drives, each actually reporting > 2.73TB. The 103 GB is interesting though, see below: > > #btrfs fi df . >> Data, RAID1: total=102.00GiB, used=101.13GiB >> Data, single: total=8.00MiB, used=0.00B >> System, RAID1: total=8.00MiB, used=16.00KiB >> System, single: total=4.00MiB, used=0.00B >> Metadata, RAID1: total=1.00GiB, used=118.70MiB >> Metadata, single: total=8.00MiB, used=0.00B >> GlobalReserve, single: total=48.00MiB, used=0.00B > > > Why am I only getting 102GB of data space? > > Google presented me a bunch of similar problems, but all of the ones I > looked at were 1) running out of either meta or data space, 2) using > devices of different sizes, or 3) needed to re-balance the pool (Normally > resulting from either 1 or 2). My problem doesn't seem to match any of > those scenarios, and my google-foo can't seem to get me any different > suggestions. > > Are there any BTRFS gurus on this list that might be able to help? > _______________________________________________ PLUG mailing list [email protected] http://lists.pdxlinux.org/mailman/listinfo/plug
