IJ22932: DISK FILLS ON FILE I/O ERRORS DURING REMOTE INSTALLATIONS

IBM Tivoli Monitoring 6.3.0 Fix Pack 7 Service Pack 6 (6.3.0.7-TIV-ITM-SP0006)

APAR status

• Closed as program error.

Error description

• Environment:
  
  
  ITM 630 FP7 TEMS
  
  ITM 630 FP7 TEMA
  
   Do you think the latest patch applied is involved? No
  
  Problem Description:
  
   If the download processing during a remote deployment of a
  package encounters a File I/O error during the writing of the
  package to the disk, then the last 4096 bytes of the download
  package are repeatedly written to the disk until the disk fills.
  Even if the disk capacity is expanded, the file system continues
  to return errors to any application that has any files currently
  opened until those files are closed and reopened.
  
  Detailed Recreation Procedure:
  
   Install ITM 630 Monitoring Server
  
  Install ITM 630 OS Agent on target endpoint
  
  Install ITM 630 R4 Agent on the target endpoint
  
  Add ITM 630 OS Agent to the TEMS Depot using addbundles command
  
  Temporarily fill the available disk space by running the command
  
  dd if=/dev/zero of=/swapfile bs=BBBB count=XXXXXXXXX
  
  where BBBB is the system block size,  XXXXXXXXX is the number of
  blocks to allocate.
  
  BBBB can be determined by executing the command
  
  lsfs -q
  
  and locate the 'block size' for the mounted drive where
  $CANDLEHOME is located
  
  XXXXXXXXX  can be determined by obtaining the number of Free
  blocks available on the $CANDLEHOME mount point using the
  command
  
  df
  
  and dividing it by BBBB.
  
  Deploy the ITM OS Agent to the endpoint using the command
  
  tacmd updateagent -n hostname:LZ -t UX -v 063007000
  
  Observe that during the downloading of the files from the TEMS
  that one of the files continues to write to the same file in a
  loop when a File I/O error occurs.
  

Local fix

Problem summary

• If during the transfer of files from the Monitoring Server to
  the endpoint a write failure occurs writing the transfered data
  buffer to the disk, the endpoint will repeatedly attempt to
  write the same data buffer to the disk until the disk is full.
  
  
  If a file write error is encountered during a remote system
  update, the failure to exit the processing on the endpoint when
  the failure occurs will result in the disk filling.
  

Problem conclusion

• If a file write error occurs during a remote system update, exit
   the write processing and return an error to the user.
  
  
  The fix for this APAR is contained in the following maintenance
  packages:
  
     | service pack | 6.3.0.7-TIV-ITM-SP0006
  

Temporary fix

• Check the disk capacity on systems where remote deployments are
  to be executed to insure that there is adequate space to
  accomodate the update.
  

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

  TEMS

• Fixed component ID

  5724C04MS

Applicable component levels