Oracle Linux: XFS Freeze Operation causes the System Hang in UEK5 (Doc ID 2691262.1)

Last updated on NOVEMBER 17, 2021

Applies to:

Symptoms

When a XFS Freeze Operation (xfs_freeze) with UEK5 it can cause a deadlock situation, we can see below trace:

2020-04-27 22:09:56.670655  Info  freeze_safe after returning from startproc : sig_handle=0 3206309.601812] RIP: 0033:0x7fcf3f6269f0
[3206309.605202] RSP: 002b:00007ffe2ef1b158 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
[3206309.610363] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fcf3f6269f0
[3206309.616282] RDX: 0000000000000180 RSI: 00007ffe2ef1b2f0 RDI: 0000000000000004
[3206309.621509] RBP: 0000000000000180 R08: 0000000000000000 R09: 0000000000000000
[3206309.626658] R10: 0000000000000021 R11: 0000000000000246 R12: 00007fcf3f6269e0
[3206309.631728] R13: 0000000000000004 R14: 00007ffe2ef1b2f0 R15: 0000000000000000
[3206309.636664] INFO: task safefreeze:13048 blocked for more than 120 seconds.
[3206309.640818]       Not tainted 4.14.35-1902.11.3.el7uek.x86_64 #2
[3206309.644218] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[3206309.649758] safefreeze      D    0 13048  12922 0x00000080
[3206309.654339] Call Trace:
[3206309.656561]  __schedule+0x2bc/0x89b
[3206309.659822]  ? __wake_up+0x13/0x15
[3206309.662911]  ? rcu_sync_enter+0x10d/0x116
[3206309.666449]  schedule+0x36/0x7c
[3206309.669454]  percpu_down_write+0xd0/0xd9
[3206309.672992]  freeze_super+0x87/0x1a2
[3206309.677427]  xfs_fs_freeze_super+0x28/0x30 [xfs]
[3206309.693520]  do_vfs_ioctl+0x1d9/0x602
[3206309.697391]  SyS_ioctl+0x79/0x84
[3206309.700994]  do_syscall_64+0x79/0x1b9
[3206309.704673]  entry_SYSCALL_64_after_hwframe+0x169/0x0

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.