Interstage Shunsaku Data Manager Troubleshooting - Microsoft(R) Windows(R) 2000/ Microsoft(R) Windows Server(TM) 2003 - - UNIX -

1.6.2 Recovering from Searcher Failovers

This section explains the following recovery procedures:

• To restore while switching to the alternative server
• To restore while waiting for a connection to the searcher

Refer to HA Functions in the Operator's Guide for more information on the failover function for the searcher.

To restore while switching to the alternative server

If a searcher failover occurs, the following messages are output to the event log or the system log (syslog) on the director server:

shn01903i: Searcher connection was broken. (192.168.10.1,33501) (director) [shund1] Shunsaku System Name=shunsaku
shn01904i: Switching to alternative searcher. (director) [shund1] Shunsaku System Name=shunsaku
shn01110i: Connected to searcher. (192.168.10.20,33501) (director) [shund1] Shunsaku System Name=shunsaku

Recovery Procedure

To return both the failed searcher and the alternative searcher back to their original configuration after a failover, restore the failed searcher then perform the following procedure:

1. Execute the shunsysstart command on the server where the searcher is located to start the restored searcher:

   shunsysstart   -n  Shunsaku system name

   When the searcher is started, the search data on the alternative searcher is automatically sent back to the restored searcher. Once this data has been resent, the alternative searcher that has been operating enters the alternative standby state to prepare for future searcher failures.

   The following operations cannot be executed while search data is being automatically resent from the alternative searcher:

   • Update operations from applications
   • The shundclear command
   • The shundresend command
   • The shundcds command
   • The shundimport command
   • The shundexport command
   • The shundbackup command
   • The shundrecover command
   • The shunsysstop command
   • The shunsyscfgedit command

This restores the system to its original configuration and restarts failover operations.

To restore while waiting for a connection to the searcher

If the number of the specified searchers that have produced abnormalities is more than the number of the degradable searchers (DegradableSearcherCnt parameter), the status will be switched to waiting for searchers to connect to and the following message is output to the system log (syslog) or the event log on the director server:

shn01906w: Searchers to operate are insufficient. Number of searchers =18 (director) [shund1] Shunsaku System Name=shunsaku

Note: The alternative searchers that have been switched are not counted in the number of searchers that have produced abnormalities.

Recovery Procedure

1. Execute the shundstate command and check whether the status of the director (State) is "WAITING":

   shundstate -s director identifier

2. Execute the shundstate command with the -p option and check whether the status of the searchers (State) is "INACTIVE". List all the searchers that have been used and the alternative servers to make checking easier.

   shundstate -s director identifier -p

3. After checking the status of all searchers in Step 2 and eliminating all the errors, execute the shunsysstart command on the search servers where the searchers are located and start the searcher:

   shunsysstart -n Shunsaku system name

4. Execute the shundstate command on all directors on which the errors occurred and wait until the status of the directors (State) is "ACTIVE".

If the degradation has been performed, operate the recovering from degradation continuously.

Refer to 1.4.1 Recovering from Errors for more information.