3742270 - LNX: HanaWorker process hang at function call nfs_file_fsync

• The HANA database server had a hang situation. 
  
  
• The OS log (/var/log/messages) includes following kernel call stack trace:
  
  

  Apr  2 02:05:39 abc01 kernel: INFO: task HanaWorker:1234567 blocked for more than 120 seconds.
  Apr  2 02:05:39 abc01 kernel:      Not tainted 4.18.0-477.104.1.el8_8.x86_64 #1
  Apr  2 02:05:39 abc01 kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.

  kernel: INFO: task HanaWorker:1234567 blocked for more than 120 seconds.
  kernel:      Not tainted 4.18.0-477.104.1.el8_8.x86_64 #1
  kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
  kernel: task:HanaWorker      state:D stack:    0 pid:1234567 ppid:654321 flags:0x00000080
  kernel: Call Trace:
  kernel: __schedule+0x2d1/0x870
  kernel: schedule+0x55/0xf0
  kernel: io_schedule+0x12/0x40
  kernel: wait_on_page_bit+0xfd/0x220
  kernel: ? filemap_fdatawait_keep_errors+0x50/0x50
  kernel: wait_on_page_writeback+0x2b/0x90 
  kernel: __filemap_fdatawait_range+0x88/0xf0
  kernel: ? __filemap_fdatawrite_range+0xd9/0x100
  kernel: file_write_and_wait_range+0x7a/0xa0
  kernel: nfs_file_fsync+0x9e/0x1a0 [nfs]
  kernel: do_fsync+0x38/0x70
  
  

  The function nfs_file_fsync is responsible for synchronizing a file’s data to a storage device over a Network File System (NFS). The above shows that the HanaWorker process is stuck waiting for this operation to complete (wait_on_page_writeback) for more than the default 120-second timeout. Then it entered an uninterruptible sleep state (state:D).

• Linux server
• HANA database

hang, HanaWorker, state:D, wait_on_page_bit, filemap_fdatawait_keep_errors, wait_on_page_writeback, file_write_and_wait_range, nfs_file_fsync , KBA , BC-OP-LNX , Linux , Problem