ENOSPC issue -- plenty of free space in metadata/data chunks

2013-08-31 Thread George Amvrosiadis
B, used=0.00 gamvrosi@c154:~$ sudo btrfs fi sh Label: none uuid: 9c632287-214b-4203-94e9-5411ad0d1fe0 Total devices 1 FS bytes used 85.86GB devid1 size 213.52GB used 89.04GB path /dev/sda3 Btrfs v0.20-rc1-253-g7854c8b-dirty -- George Amvrosiadis, B.Sc., Ph.D. Candidate Grad

Re: filebench varmail + scrubber = btrfs_update_root bug

2013-07-16 Thread George Amvrosiadis
> > > > unable to update root key 18446744073709551610 132 5 > Is (objectid = -6 = BTRFS_TREE_LOG_OBJECTID, type = 132 = > BTRFS_ROOT_ITEM_KEY, offset = 5) > Stefan, thanks for clarifying that 18446744073709551606 is a valid objectid. I somehow failed to notice those defini

filebench varmail + scrubber = btrfs_update_root bug

2013-07-14 Thread George Amvrosiadis
2558.688239] [] system_call_fastpath+0x16/0x1b [ 2558.688245] ---[ end trace 1cf4adc709e16e0b ]--- And that root key (18446744073709551610) looks very suspicious. Any ideas? Thanks, -- George Amvrosiadis, B.Sc., Ph.D. Student Graduate Students' Union Representative Computer Science, Unive

btrfs scrubber: clarification on scrub_chunk

2013-04-13 Thread George Amvrosiadis
} 2540 } Could someone please clarify line 2533? My intuition from the code so far was that this should be "map->stripes[i].physical == dev_offset + i * stripe_size", otherwise the condition would fail for all but the first stripe of every chunk (effectively scrubbing only