P101035: LIM RESTART OLD LIM DOESN'T EXIT, NEW LIM CANNOT START

APAR status

• Closed as program error.

Error description

• 1) At about Mar 19 04:06:31, or a little before this time point,
  a
  "lsadmin lim restart" was executed on master houcy1-n-sp099a02.
  2) But as master's load was high, the original lim had not
  exited, and
  not released the port. At meanwhile, the new lim failed to
  startup due
  to the port unavailable. So we see the messages in lim log as
  below:
  
  Mar 19 04:06:31 2015 46069 3 1.2.7 initSock():
  chanServSocketExt_(). A
  socket operation has failed on the configured UDP port <7869> on
  host
  <houcy1-n-sp099a02>. Reason: <Address already in use>. Fatal
  error.
  Either change the port number in lsf.conf (LSF_LIM_PORT) or
  terminate
  the other process that is bound to the port.
  Mar 19 04:06:31 2015 46069 3 1.2.7 initSock: LIM has exited due
  to a
  fatal error.
  
  3) As the lim was in abnormal status on master, a failover
  occurred.
  During the failover, a job submission was committed, but it
  failed with
  messages as below:
  LSF is down. Please wait ...
  
  Connection refused by server. Job not submitted.
  

Local fix

• n/a
  

Problem summary

• When the system is too busy to release a port, it will cause the
   lim/res restart to fail because the socket failed to
  initialize.
  
  This fix introduces the following parameters in lsf.conf to
  control the retry behavior:
  

Problem conclusion

• fix it
  

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

  LSF STD LEGACY

• Fixed component ID

  5725G8206

Applicable component levels

• R911 PSY

     UP