This blog contains experience gained over the years of implementing (and de-implementing) large scale IT applications/software.

HANA – Emergency Shutdown

Darryl Griffiths
For testing purposes, you can perform an emergency shutdown of the HANA indexserver forcing it to simply crash out.
This will require a DB recovery (automatic) upon restart!

Use the hdbcons utility (as <sid>adm) to first activate the crash functionality, then invoke the crash:

> hdbcons
> crash activate
> crash emergencyshutdown

Inside the indexserver trace file (from diagnosis view in HANA Studio), you will be able to see the following text towards the end of the file “ExternalCommandHandler.cpp (00958)  NOTE: INTENTIONAL CRASH: errorExitWithoutCrashdump“.

Once the restart is complete, you can verify that recovery was required from the indexserver trace file:

“LoggerImpl.cpp(00933) : Replayed 130944B (0MB) of log in 17.4729 seconds; 0.00714693MB/s; max known TID=50175“

2 thoughts on HANA – Emergency Shutdown

  1. Nice blog, how do we totally crash HANA System ? I wanted to test the Disaster Recovery scenario crashing the SAP HANA System.

    Thanks
    Srikar

    Reply

  2. Using the above method does totally crash the HANA system.
    The other methods you could use:

    – "kill -9" on the indexserver process.
    – Drop the HBA connection to the SAN.
    – Switch off the power.
    – Use a utility program to consume all available O/S memory to try and force a HANA OOM.
    – Physically pull a disk/SSD out.
    – Pull a memory DIMM out.
    – Use "cpio" or "dd" to overwrite data files and/or log files.
    – In SUSE issue a "panic".
    – Switch off the air conditioning to see when/if the server will overheat.
    – Switch off ntpd and then reset the time in the O/S, to a previous date/time (this may not do anything).

    Regards,

    Darryl

    Reply

Add Your Comment

* Indicates Required Field

Your email address will not be published.

*