|
After racer is finished, I saw some dd threads can not be stopped
17:10:58:Stopping client client-32vm1.lab.whamcloud.com /mnt/lustre2 opts:
17:10:59:COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
17:10:59:dd 1756 root 1w REG 1273,181606 49091584 144115205306079415 /mnt/lustre2/racer/19
17:11:01:dd 9219 root 1w REG 1273,181606 8193024 144115205306056725 /mnt/lustre/racer/15
17:11:01:dd 9224 root 1w REG 1273,181606 8193024 144115205306056725 /mnt/lustre2/racer/15
17:11:02:dd 9225 root 1w REG 1273,181606 8193024 144115205306056725 /mnt/lustre2/racer/15
17:11:02:dd 11097 root 1w REG 1273,181606 245867520 144115205255725671 /mnt/lustre/racer/13
17:11:02:Stopping client client-32vm2.lab.whamcloud.com /mnt/lustre2 opts:
17:11:02:/mnt/lustre2 is still busy, wait one second
17:11:02:COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
17:11:02:dd 12954 root 1w REG 1273,181606 119452672 144115205255740618 /mnt/lustre/racer/11
17:11:03:dd 13442 root 1w REG 1273,181606 116897792 144115205289288929 /mnt/lustre2/racer/0
17:11:03:dd 18335 root 1w REG 1273,181606 245867520 144115205255725671 /mnt/lustre/racer/10
17:11:03:dd 22305 root 1w REG 1273,181606 65881088 144115205272518234 /mnt/lustre2/racer/15 (deleted)
17:11:03:/mnt/lustre2 is still busy, wait one second
17:11:03:/mnt/lustre2 is still busy, wait one second
17:11:03:/mnt/lustre2 is still busy, wait one second
17:11:03:/mnt/lustre2 is still busy, wait one second
17:11:05:/mnt/lustre2 is still busy, wait one second
17:11:05:/mnt/lustre2 is still busy, wait one second
17:11:05:/mnt/lustre2 is still busy, wait one second
17:11:05:/mnt/lustre2 is still busy, wait one second
17:11:05:/mnt/lustre2 is still busy, wait one second
17:11:05:/mnt/lustre2 is still busy, wait one second
.....
It happened to me that these threads were doing single page RPC to flush the dirty data to the server, as I investigated the log before. Unfortunately, I do not have debug log right now.
|