DB2 - Problem description
Problem IT24866 | Status: Closed |
IN A HADR HA(TSA) ENVIRONMENT, IN THE EVENT OF A PRIMARY SERVER FAILURE, THE STANDBY MAY NOT BE REINTEGRATED ON INSTANCE STARTUP | |
product: | |
DB2 FOR LUW / DB2FORLUW / B10 - DB2 | |
Problem description: | |
This issue is due to a very small timing hole in the code. The Reintegration flag for the database is created at around the same time that we check for its existence in the db2V111_start.ksh script. If the check for the flag in the db2V111_start.ksh script happens before its creation, then the HADR database will be activated instead of reintegrated as the new standby via the "START HADR AS STANDBY" command. The following logs from the db2V111_start.ksh script will be seen if this issue is encountered: Jun 22 09:39:34 ****-**-vm0047 db2V111_start.ksh[4130]: Entered /usr/sbin/rsct/sapolicies/db2/db2V111_start.ksh, db2inst1, 0 Jun 22 09:39:52 ****-**-vm0047 db2V111_start.ksh[4130]: Running HADR monitor command for database SAMPLE on host ****-**-vm0046 Jun 22 09:39:53 ****-**-vm0047 db2V111_start.ksh[4130]: su - db2inst1 -c db2 activate database SAMPLE Jun 22 09:39:58 ****-**-vm0047 db2V111_start.ksh[4130]: su - db2inst1 -c db2 activate database SAMPLE In the case that the host ****-**-vm0047 was hosting the primary role for db SAMPLE and was then rebooted. The db2V111_start.ksh script should reintegrate the database as a standby instead of activating it on start up. If the database is activated in this case, then this issue is being hit. | |
Problem Summary: | |
**************************************************************** * USERS AFFECTED: * * Automated HADR/TSA environments * **************************************************************** * PROBLEM DESCRIPTION: * * See Error Description * **************************************************************** * RECOMMENDATION: * * Upgrade to Db2 V11.1 Mod 1 Fix Pack 1 * **************************************************************** | |
Local Fix: | |
Adding the following 1 second sleep in the db2V111_start.ksh will fix this issue: # run HADR monitor script on other node to ensure Reintegration flag is created before # we check for it runRemoteHADRMonitor sleep 1 # comment following line to not restart or activate database at instance start time activateDatabase | |
Solution | |
Workaround | |
not known / see Local fix | |
BUG-Tracking | |
forerunner : IT15855 follow-up : | |
Timestamps | |
Date - problem reported : Date - problem closed : Date - last modified : | 25.04.2018 25.04.2018 25.04.2018 |
Problem solved at the following versions (IBM BugInfos) | |
Problem solved according to the fixlist(s) of the following version(s) |