IJ04977: ERROR RELATED TO NODE TIMEOUT DURING SNAPSHOT RESTORE

PowerHA-7.2.1 SP3

APAR status

• Closed as program error.

Error description

• Environment details: standard cluster Cluster type:  Number of
  nodes: 2 Nodes details : r1r2m1p34 (172.19.67.44) , r1r2m1p35
  (172.19.67.45) root/Ar1cent@2 HMC details (If nodes are not
  public LPARs): 172.19.69.5
  
  PowerHA version:  Note: Specify pre and post migration PowerHA
  versions if issue is related to migration.  Premigration PHA
  720 sp5 postmigration PHA 7.2.1 SP3
  
  AIX version:  oslevel is same 7100-04-03-1642
  
  Preconditions (If any):  Problem Statement:  1) Cluster created
  and is stable for 720 sp5 .  2) Started migration task from 720
  sp5 -> 721 sp3 rolling migration 3) Post migration cluster is
  stable 4) took cluster snapshot of latest post migrated cluster
  721 sp3 5) deleted the cluster succesfully 6) trying to restore
  the snapshot of the earlier 721 sp3 cluster scuessfull
  (snap721) 7) During verify and sync getting error stated below
  
  ERROR: node_timeout must be greater than network_fdt by at
  least 10 seconds. Correct these values and try synchronization
  again.
  
  cldare: Error detected during synchronization.
  
  Steps to reproduce: As above mentioned Problematic behavior:
  verify sync failed Expected behavior: after snapshot
  restoration verify sync should pass References (If applicable):
  Publib infocenter link or old defect details.  % of
  reproducibility:  Logs location or list of logs attached:
  
  (0) root @ r1r2m1p34: /
  
      /usr/sbin/clctrl -tune -L network_fdt NAME DEF MIN MAX UNIT
      SCOPE ENTITY_NAME(UUID) CUR
      ------------------------------------------------------------
  --------------------
      network_fdt 0 0 590000 milliseconds c
      r1r2m1p34_cluster(a7299d1c-1193-11e8-8042-caaa6459140c)
      20000
      ------------------------------------------------------------
  --------------------
  
  (0) root @ r1r2m1p34: /
  
      /usr/sbin/clctrl -tune -L node_timeout NAME DEF MIN MAX
      UNIT SCOPE ENTITY_NAME(UUID) CUR
      ------------------------------------------------------------
      node_timeout 20000 10000 600000 milliseconds c n
      r1r2m1p34_cluster(a7299d1c-1193-11e8-8042-caaa6459140c)
      30000
      ------------------------------------------------------------
  --------------------
  

Local fix

Problem summary

• Getting below error during pha 721 sp5 snapshot restoration
  ERROR: node_timeout must be greater than network_fdt by at
  least 10 sec This is because of mismatch of odm values before
  and after restoring snapshot before snapshot: node_timeout=20
  network_fdt=0 after snapshot: node_timeout=20 network_fdt=20,
  here both the values are same, which is not meeting the below
  condition node_timeout - network_fdt = 10
  

Problem conclusion

• The default values of node_timeout is 30ms and network_fdt is
  20. As per the existing code we are setting the network_fdt and
  node_timeout to default values if the odm values are 0. But
  node_timeout value is not 0 thats why it is not setting to the
  default values instead setting it to the existing odm value
  which is 20. To fix this, Make the code changes such that
  node_timeout should set to default value which is 30 when the
  value is 20 instead of 0
  

Temporary fix

Comments

APAR Information

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

  IJ04976

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

Fix information

• Fixed component name

  POWERHA SYSMIR

• Fixed component ID

  5765H3900

Applicable component levels