Peer CSR

Peer CSR is skipped unless it is necessary to perform group restart on behalf of non-starting members. Peer CSR is not performed on non-starting members that are normally quiesced.

A restarting member can select an inactive member on which to perform peer initialization and peer CSR:

If the SCA is lost, the restarting member rebuilds SCA information from the information contained in the non-starting member's logs.
If the lock structure is lost, the restarting member reacquires page set and partition P-locks (as retained locks) for the non-starting member. Those locks are now retained locks.

During peer CSR, you see messages similar to the following messages. (The phrases in parentheses are not part of the output.)

 DSNR023I @DB3ADB2 DSNRRGRC GROUP RESTART INITIATED TO
 RECOVER THE SCA FOR GROUP MEMBER DB1A
 DSNR003I @DB3ADB2 RESTART...PRIOR CHECKPOINT RBA=00000201CC4E
 DSNR004I @DB3ADB2 RESTART...UR STATUS COUNTS
 IN COMMIT=0, INDOUBT=0, INFLIGHT=0, IN ABORT=0
 DSNR024I @DB3ADB2 DSNRRGRC GROUP RESTART COMPLETED TO
 RECOVER THE SCA FOR GROUP MEMBER DB1A
 (End of peer current status rebuild for member DB1A)
 
 DSNR023I @DB3ADB2 DSNRRGRC GROUP RESTART INITIATED TO
 RECOVER THE SCA FOR GROUP MEMBER DB2A
 DSNR003I @DB3ADB2 RESTART...PRIOR CHECKPOINT RBA=000000009C4E
 DSNR004I @DB3ADB2 RESTART...UR STATUS COUNTS
 IN COMMIT=0, INDOUBT=0, INFLIGHT=0, IN ABORT=0
 DSNR024I @DB3ADB2 DSNRRGRC GROUP RESTART COMPLETED TO
 RECOVER THE SCA FOR GROUP MEMBER DB2A
 (End of peer current status rebuild for DB2A)
 
 DSNR022I @DB3ADB2 DSNRRGRC DB2 SUBSYSTEM HAS
 COMPLETED GROUP RESTART FOR PEER MEMBERS
 (End of peer processing)

When all members have completed CSR—either by performing it on their own or by having a peer perform it for them—the SCA has been rebuilt, and page set and partition P-locks have been reacquired.