[Samba] task "smbd" blocked for more than x seconds (followed by call trace)

[Jeremy Allison]

On Mon, Apr 01, 2024 at 08:52:54AM -0700, Alex via samba wrote:
>Hi,
>
>I have a samba 4.15.13 file server running on Ubuntu 20.04.6, fully
>updated/patched. The samba shared folders are stored on an XFS formatted
>drive.
>Intermittently, usually after 1 or 2 weeks of normal operation, browsing
>the shared drive will freeze Windows Explorer, and when I look at the file
>server console, I see the messages below over and over again:
>
>[2955729.823684] INFO: task smbd:2888777 blocked for more than 120 seconds.
>[2955729.824540]       Not tainted 5.4.0-172-generic #190-Ubuntu
>[2955729.824989] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs"
>disables this message.
>[2955729.825549] smbd            D    0 2888777   1737 0x00000080
>[2955729.825565] Call Trace:
>[2955729.825674]  __schedule+0x2e3/0x740
>[2955729.825689]  schedule+0x42/0xb0
>[2955729.825698]  rwsem_down_read_slowpath+0x16c/0x4a0
>[2955729.825760]  __down_read+0x6b/0x80
>[2955729.825769]  __percpu_down_read+0x54/0x80
>[2955729.825802]  __sb_start_write+0x79/0x80
>[2955729.825848]  mnt_want_write+0x24/0x60
>[2955729.825866]  do_last+0x8ea/0x900
>[2955729.825877]  ? __alloc_file+0x94/0x110
>[2955729.825890]  path_openat+0x8d/0x290
>[2955729.825904]  do_filp_open+0x91/0x100
>[2955729.825923]  ? __alloc_fd+0x46/0x150
>[2955729.825954]  do_sys_open+0x17e/0x290
>[2955729.825991]  ? __audit_syscall_exit+0x233/0x290
>[2955729.826005]  __x64_sys_openat+0x20/0x30
>[2955729.826036]  do_syscall_64+0x57/0x190
>[2955729.826055]  entry_SYSCALL_64_after_hwframe+0x5c/0xc1
>[2955729.826085] RIP: 0033:0x7fac5d210163
>[2955729.826098] Code: Bad RIP value.
>[2955729.826103] RSP: 002b:00007ffec466ae50 EFLAGS: 00000293 ORIG_RAX:
>0000000000000101
>[2955729.826119] RAX: ffffffffffffffda RBX: 0000555ed448f100 RCX:
>00007fac5d210163
>[2955729.826124] RDX: 00000000000208c2 RSI: 0000555ed441fd20 RDI:
>00000000ffffff9c
>[2955729.826128] RBP: 0000555ed441fd20 R08: 0000000000000000 R09:
>00000000000001e4
>[2955729.826133] R10: 00000000000001e4 R11: 0000000000000293 R12:
>00000000000208c2
>[2955729.826138] R13: 00000000000001e4 R14: 0000555ed43c7b40 R15:
>0000555ed43c7b40
>[2955850.658135] INFO: task smbd:2888777 blocked for more than 241 seconds.
>[2955850.659195]       Not tainted 5.4.0-172-generic #190-Ubuntu
>[2955850.659800] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs"
>disables this message.

You have a bad or corrupted disk. Nothing smbd does
causes kernel hangs.