IT33425: VSNAP KERNEL PANIC USING VERSION 10.1.5 DURING ZFS POOL SCRUB

APAR status

• Closed as program error.

Error description

• A vSnap host kernel panic may occur during a zfs pool scrub.
  As a result the vSnap host may repeatedly crash and reboot.
  In the crash dump (vmcore-dmesg.txt) under /var/crash,
  the following call trace is seen at the end of the file:
  
  Call Trace:
  
  rs_get_start+0xc/0x20 [zfs]
   scan_io_queue_fetch_ext+0x52/0x90
  [zfs]
   scan_io_queues_run_one+0xde/0x250 [zfs]
  
  taskq_thread+0x213/0x3d0 [spl]
   ? wake_up_q+0x70/0x70
  
  kthread+0xf8/0x130
   ? task_done+0x80/0x80 [spl]
   ?
  kthread_bind+0x10/0x10
   ret_from_fork+0x35/0x40
  
  IBM Spectrum
  Protect Plus Versions Affected:
  IBM Spectrum Protect Plus
  10.1.5
  
  | MDVREGR 10.1.5-5737SPLUS |
  
  Initial Impact:
  Medium
  
  Additional Keywords: SPP, SPPLUS, TS003811628, crash,
  zfs
  

Local fix

• 1. Find the pool ID with the command : zpool status
  2. Cancel
  the scrub using the command : zpool scrub -s vpool<ID>
  The
  scrub can be resumed once updated to the fixing version.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * IBM Spectrum Protect Plus level 10.1.5                       *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * See Error Description                                        *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * Apply fixing level when available. This problem is currently *
  * projected to be fixed in IBM Spectrum Protect Plus level     *
  * 10.1.5.2218 and 10.1.6. Note that this is subject to change  *
  * at the discretion of IBM.                                    *
  ****************************************************************
  

Problem conclusion

• The kernel module for the filesystem contained an incorrect
  check for a null pointer during pool scrub. This resulted in a
  kernel panic and system crash. The problem has been resolved in
  an updated version of the kernel module which has been
  incorporated into a newer version of vSnap.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  SP PLUS

• Fixed component ID

  5737SPLUS

Applicable component levels