OLVM: I/O issue if direct LUN drived by VirtIO-SCSI with SCSI Pass-Through enabled
(Doc ID 2788001.1)
Last updated on OCTOBER 05, 2023
Applies to:
Linux OS - Version Oracle Linux 7.9 with Unbreakable Enterprise Kernel [5.4.17] and laterLinux x86-64
Symptoms
Symptoms could be either one of below or some/all combined:
- VM becomes paused with error message "VM has been paused due to unknown storage error"
- VM hangs under high I/O activity
- With a Device Mapper Multipath device mapped via VirtIO-SCSI with SCSI Pass-Through enabled, I/O only go through one path constantly
- Below call trace found inside VM:
Jun 3 06:02:36 olvm-vm kernel: INFO: task ora_task:1905 blocked for more than 122 seconds.
Jun 3 06:02:36 olvm-vm kernel: Tainted: P O 5.4.17-2102.200.13.el7uek.x86_64 #2
Jun 3 06:02:36 olvm-vm kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Jun 3 06:02:36 olvm-vm kernel: ora_task D 0 1905 1 0x00000080
Jun 3 06:02:36 olvm-vm kernel: Call Trace:
Jun 3 06:02:36 olvm-vm kernel: __schedule+0x2bc/0x599
Jun 3 06:02:36 olvm-vm kernel: schedule+0x3e/0x9a
Jun 3 06:02:36 olvm-vm kernel: io_schedule+0x16/0x39
Jun 3 06:02:36 olvm-vm kernel: __blkdev_direct_IO_simple+0x20f/0x2e1
Jun 3 06:02:36 olvm-vm kernel: ? blkdev_iopoll+0x40/0x31
Jun 3 06:02:36 olvm-vm kernel: blkdev_direct_IO+0x42b/0x493
Jun 3 06:02:36 olvm-vm kernel: ? current_time+0x41/0x78
Jun 3 06:02:36 olvm-vm kernel: ? atime_needs_update+0x6c/0xca
Jun 3 06:02:36 olvm-vm kernel: ? touch_atime+0x33/0xca
Jun 3 06:02:36 olvm-vm kernel: generic_file_read_iter+0x8d/0x13a
Jun 3 06:02:36 olvm-vm kernel: blkdev_read_iter+0x37/0x3a
Jun 3 06:02:36 olvm-vm kernel: new_sync_read+0x11a/0x1be
Jun 3 06:02:36 olvm-vm kernel: __vfs_read+0x29/0x34
Jun 3 06:02:36 olvm-vm kernel: vfs_read+0x8e/0x138
Jun 3 06:02:36 olvm-vm kernel: ksys_pread64+0x69/0x94
Jun 3 06:02:36 olvm-vm kernel: __x64_sys_pread64+0x1e/0x20
Jun 3 06:02:36 olvm-vm kernel: do_syscall_64+0x60/0x1d9
Jun 3 06:02:36 olvm-vm kernel: entry_SYSCALL_64_after_hwframe+0x170/0x0
Jun 3 06:02:36 olvm-vm kernel: RIP: 0033:0x7f5db2ce9fa3
Jun 3 06:02:36 olvm-vm kernel: Code: Bad RIP value.
Jun 3 06:02:36 olvm-vm kernel: RSP: 002b:00007ffdcb704728 EFLAGS: 00000246 ORIG_RAX: 0000000000000011
Jun 3 06:02:36 olvm-vm kernel: RAX: ffffffffffffffda RBX: 00007ffdcb70a7a8 RCX: 00007f5db2ce9fa3
Jun 3 06:02:36 olvm-vm kernel: RDX: 0000000000004000 RSI: 00007f5db1771000 RDI: 0000000000000100
Jun 3 06:02:36 olvm-vm kernel: RBP: 00007ffdcb709130 R08: 00007ffdcb704710 R09: 00007ffdcb7e20b0
Jun 3 06:02:36 olvm-vm kernel: R10: 0000000001804000 R11: 0000000000000246 R12: 0000000000000200
Jun 3 06:02:36 olvm-vm kernel: R13: 0000026a60d85478 R14: 0000000000000000 R15: 00007f5db17e1000
Changes
By adding a direct LUN drived by VirtIO-SCSI, the "Enable SCSI Pass-Through" is checked by default, which should not be as it is under technical preview at the time of writing.
Cause
To view full details, sign in with your My Oracle Support account. |
|
Don't have a My Oracle Support account? Click to get started! |
In this Document
Symptoms |
Changes |
Cause |
Solution |
References |