I am using mdtest benchmarking tool to remove 7 million files of 1KB. After some time, the mdtest process gets blocked. Dmesg shows the following --------------------->8------------ May 10 18:49:00 machine1 kernel: INFO: task xfssyncd/dm-3:8156 blocked for more than 120 seconds. May 10 18:49:00 machine1 kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. May 10 18:49:00 machine1 kernel: xfssyncd/dm-3 D 0000000000000002 0 8156 2 0x00000080 May 10 18:49:00 machine1 kernel: ffff88185932bc70 0000000000000046 ffff88185e58a400 ffff8800282566e8 May 10 18:49:00 machine1 kernel: 00000000000176f6 ffff88185fb44080 ffff880c6100aaa0 ffffffff8160b400 May 10 18:49:00 machine1 kernel: ffff88185fb44638 ffff88185932bfd8 000000000000fb88 ffff88185fb44638 May 10 18:49:00 machine1 kernel: Call Trace: May 10 18:49:00 machine1 kernel: [<ffffffffa056c492>] xlog_wait+0x72/0x90 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffff81060250>] ? default_wake_function+0x0/0x20 May 10 18:49:00 machine1 kernel: [<ffffffffa056e08b>] xlog_grant_log_space+0x3ab/0x520 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffffa0582d1a>] ? kmem_zone_zalloc+0x3a/0x50 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffff8127f3ac>] ? random32+0x1c/0x20 May 10 18:49:00 machine1 kernel: [<ffffffffa057d201>] ? xfs_trans_ail_push+0x21/0x80 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffffa056e2e6>] xfs_log_reserve+0xe6/0x140 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffffa057bad0>] xfs_trans_reserve+0xa0/0x210 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffffa055f0a3>] xfs_fs_log_dummy+0x43/0x90 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffffa056c9c4>] ? xfs_log_need_covered+0x94/0xd0 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffffa0591831>] xfs_sync_worker+0x81/0x90 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffffa059171e>] xfssyncd+0x17e/0x210 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffffa05915a0>] ? xfssyncd+0x0/0x210 [xfs] May 10 18:49:00 machine1 kernel: [<ffffffff81091d66>] kthread+0x96/0xa0 May 10 18:49:00 machine1 kernel: [<ffffffff8100c14a>] child_rip+0xa/0x20 May 10 18:49:00 machine1 kernel: [<ffffffff81091cd0>] ? kthread+0x0/0xa0 May 10 18:49:00 machine1 kernel: [<ffffffff8100c140>] ? child_rip+0x0/0x20 --------------------->8------------ The XFS filesystem size is 14TB (LVM), kernel version 2.6.32 Machine has a memory of 96GB Please let me know if this is a known issue and have been already fixed in higher kernel version. -- Subranshu _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
