Index server cannot be started for OOM after stop all servers

[Former Member]

     Recently, I stopped all HANA servers in test environment, but the index server cannot be started after stop all servers, it is only statistics server started, however others server cannot be started, and the HANA server showed error message, the index server cannot be started.

It maybe has various situation for this issue, I will share my experience for OOM (out of memory) situation.

Firstly, you can find your HANA server trace file for each server, it is at /usr/sap/<SID>/<SID><instance number>/<hostname>/trace on HANA server side.

You can find indexserver_<hostname>.port(default 30003).rtedump.<timestemp>.oom.trc file in trace folder, please check below figure.

You can find the "information about current out of memory situation" message in this trace file.

Next step, we need to check global.ini file to check global allocation limit for memory.

You can find local global.ini at /usr/sap/<sid>/<sid><instance number>/<hostname>/ on HANA server, you need to check this limit value is normal or not based on your HANA server physical memory size and allow max value of HANA memory licence, you need to revise  and increase memory based on memory allocation concept, its value <= 90% physical memory as well, because others memory will be used for OS and others process.

If you can not increase allocation memory based on current physical memory or allow max value of HANA memory licence, then you need to buy physical memory or new max value of HANA memory licence to increase allocation memory for HANA server.

Hope this can help you, if you have the similar issue.

Regards,

Jerry