Recently the Operating System firmware of Hana DB System with multiple instance has been upgraded. Therefore when logged in to OS level, running a command 'HBD stop' to stop the HANA DB system takes more time to stop in comparison to time taken before upgradation.
The command 'HDB stop" on single host is a soft stop. The system stops after all running statments gets finished or specified number of seconds <###> timeout is reached.
Please do as following to resolve this issue:
The systems stops afted specified timeout is reached or all running statments get finished.
The following points help to trace the HDB stop command. This will help pointing to the service in the case of delay, as follows:
After executing 'HDB stop' you will see the entries in the daemon<host>.3####.###.trc file
- Goto SAP HANA Studio, logon as SYSTEM user.
- Then Open Administration Editor in diagnosis mode.
- Open Diagnosis Files tab then open most recent daemon trace file
Look for, after this see the runlevels are stopped in reverse order.
Look for entries:
- got shutdown event (stop)
- stop process hdb<service> with pid #####
- process hdb<service> with pid ##### exited normally with status 0
If the timeout limit is reached look for 'got shutdown event (quit)'.
Please check if a particular thread was not in a state where it could be immediately terminated after executing command 'HDB stop'.