VM66093: VSWITCH CONNECTIVITY LOSS AFTER HEAVY TRAFFIC TO UPLINK PORT

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• After heavy outbound traffic out the VSWITCH UPLINK port, guests
  lose connectivity until the VSWITCH RDEV connection is reset.
  VSWITCH guest connectivity is not affected. Resetting the guest
  interface does not restore connectivity. VSWITCH disconnect
  and connect sequence (or VSWITCH FAILOVER) will restore guest
  connectivity.
  

Local fix

• Reset the uplink connection with the SET VSWITCH command
  until this PTF is available.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: VSWITCH users may experience connectivity    *
  *                 loss after spikes in outbound traffic        *
  *                 through the UPLINK port.                     *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  ****************************************************************
  * RECOMMENDATION: APPLY PTF                                    *
  ****************************************************************
  If a spike in outbound VSWITCH traffic causes the pending queue
  for the UPLINK port to overflow, the logic to discard those
  messages fails to operate correctly (leaving the messages in
  place to accumulate). New outbound messages encounter the same
  error, so coupled guests are no longer able to transmit to
  an external host.
  The queue can be cleared by a failover, or by a VSWITCH
  disconnect and connect sequence. This allows the affected
  hosts to restore communication across the uplink port.
  

Problem conclusion

• Entry point HCPSWASP directs outbound messages to the UPLINK
  interface. It includes logic to detect an overflow of messages
  pending for that UPLINK port. When too many messages are in
  that queue, an error in the register specification causes CP to
  clear the queue from the source device extension instead of the
  destination (UPLINK) device extension block. That error is
  corrected with this update, so the UPLINK device queue is
  properly cleared and prepared to handle subsequent outbound
  network messages. This allows the affected hosts to process
  new outbound network messages. The guests should automatically
  recover from the overflow (for example, the TCP/IP protocol
  will retransmit as necessary).
  

Temporary fix

• FOR RELEASE VM/ESA CP/ESA R640 :
  PREREQ: VM66037
  CO-REQ: NONE
  IF-REQ: NONE
  

Comments

APAR Information

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

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

  UM35249

Modules/Macros

• HCPSWA
  

Fix information

• Fixed component name

  VM CP

• Fixed component ID

  568411202

Applicable component levels

• R540 PSN

     UP

• R630 PSN

     UP

• R640 PSY UM35249

     UP18/04/24 P 1801 ¢

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.