RDSync consuming lot of RAM on HANA server

[Former Member]

Dear Experts,

We have installed RDSync on HANA server and the RDSync service is consuming a lot of RAM on the same with just 2-3 scripts created for testing. Is there a way to cut down the consumption.

[Vlad]

Can you please be more precise what the scripts are?
Do they select * from a table?
How many concurrent sync do you have?
How often?
Average response time and response size?
How many records does each table have?
The average row width (bytes). E.g. you are synchronizing boolean values, or binary data.
The more info you give us, the more answers we will write.

[Former Member]

Yes it is to select from a table. And the table has hardly 10-15 records. The sync is not scheduled. I only do it manually from SQL Anywhere when required. There are no boolean values or binary data. It is as good as a fresh instance.

[Former Member]

How much memory is being consumed by HANA and RDSync? What is the RAM size of the host?

To see the memory tracked by RDSync server directly, you can add the "-ppv 60" option the extra_options in the <sapmnt>/global/hdb/custom/config/DB_RJY/rdsyncserver.ini file, where <sapmnt> is your HANA install directory, then restart RDSync server. From then on, every 60 seconds you will see periodic values in the rdsync*.trc file. The value to look for is in the "PERIODIC: TRACKED_MEMORY:" entry. This value represents the memory explicitly tracked by RDSync server. This value will NOT include memory used by the HANA ODBC driver nor the memory used by the JRE (Java Runtime Environment). The TRACKED_MEMORY value is usually much lower than the actual RDSync server memory usage reported by the OS. It is usually the JRE that accounts for the bulk of the difference between the TRACKED_MEMORY value and the memory use shown for the RDSync server process as a whole.

• Russ