Michał Sawicz
2013-Jun-11 18:42 UTC
btrfs-transacti:1014 blocked for more than 120 seconds.
Hey, I''ve a 2x4TB RAID1 setup with btrfs on kernel 3.8.0. Under high I/O load (BackupPC dump or writing a large file over gigabit) I get messages in syslog such as the one mentioned in the subject. The full non-logcheck-ignored log is under [1]. A BackupPC dump between the same exact machines onto a 2TB ext4 volume take 90 minutes on average, the process on the btrfs volume took 465 minutes... BackupPC uses hard links for unchanged files between backups, if that''s relevant. When copying a large file over gigabit, I''m getting around 50MB/s burst speed for about 30s, followed by a minute of 0B/s when btrfs-submit and btrfs-transaction are at 99% IO in iotop but minimal READ/WRITE. Is there anything I can do to relieve that? If more details can help, I''ll gladly provide them. [1] http://pastebin.ubuntu.com/5755800/ Thanks, -- Michał (Saviq) Sawicz <michal@sawicz.net>
Michał Sawicz
2013-Jun-11 18:50 UTC
btrfs-transacti:1014 blocked for more than 120 seconds.
Hey, I''ve a 2x4TB RAID1 setup with btrfs on kernel 3.8.0. Under high I/O load (BackupPC dump or writing a large file over gigabit) I get messages in syslog such as the one mentioned in the subject. The full non-logcheck-ignored log is under [1]. A BackupPC dump between the same exact machines onto a 2TB ext4 volume take 90 minutes on average, the process on the btrfs volume took 465 minutes... BackupPC uses hard links for unchanged files between backups, if that''s relevant. When copying a large file over gigabit, I''m getting around 50MB/s burst speed for about 30s, followed by a minute of 0B/s when btrfs-submit and btrfs-transaction are at 99% IO in iotop but minimal READ/WRITE. Is there anything I can do to relieve that? If more details can help, I''ll gladly provide them. [1] http://pastebin.ubuntu.com/5755800/ Thanks, -- Michał (Saviq) Sawicz <michal@sawicz.net>