Hana server management best practices

[Former Member]

I am new to SAP Hana DB and am doing maintenance of its server as well as using predictive analysis over the existing data. However, so far the database has been frustrating to use as the machine on which the database was installed repeatedly ran out of space and started throwing errors on querying:

SAP DBTech JDBC: [3584]: distributed SQL error: [9] Memory allocation failed (com.sap.db.jdbc.exceptions.JDBCDriverException: SAP DBTech JDBC: [3584]: distributed SQL error: [9] Memory allocation failed)

I had stopped the tracefile collection and removed the old contents from the log folder which solved this problem. This worked fine for few days but today I received another baffling error:

Could not execute '<MY QUERY>' in 28.080 seconds . SAP DBTech JDBC: [2048]: column store error: search table error:  [9] Memory allocation failed

Also, the tracefiles and log files are continuously getting backed up (unless stopped explicitly) and I had deleted these backups last time. Is there some elegant mechanism to specify the backup strategy to remove old backups continually?

This is a live server where data is increasing continuously, but not at this velocity, which is consuming multiple GBs of data in few weeks.

The system statistics are fine where the memory and disk space is still showing underutilized resources, however I may be wrong here. Is there any other way of managing the server apart from changing the configuration ini for different services within the hana studio.

Thanks,

sumit

[Former Member · Accepted solution]

There are a whole load of red flags in your post.

- You show an 80GB instance and say it is productive. There is no 80GB supported HANA System so this must be unsupported hardware.

- Your memory usage is very high. You need to find out why that is. My guess is you overfilled the system. You should only have 50% used by data and 50% left for queries.

If this is a productive system then I recommend getting a seasoned HANA consultant to take a look. There may be some simple reasons for your problems.

In any case if you have problems with a productive system then you should consider raising a support call.

John

[former_member184768]

Hi Sumit,

Such errors are dependent on multiple factors.

• How the model designed on which the query is based
• How big is the requested data size. Are there any filters defined
• How many table join operations are performed
• How many column join operations are performed.

Basically it comes to application level tuning and needs additional information on the model and query pattern.

Can you please provide more information on the same.

Regards,

Ravi

[Former Member]

Hi Ravindra,

Unfortunately, all queries are failing at the moment.

Even if the table has few fields and I am explicitly limiting the rows to be fetched by the queries, it is failing.

Tables with multiple joins and unrestricted record were getting selected earlier with this database.

Thanks,

sumit

[former_member184768]

Hi Sumit,

Can you please check if your Delta merge has been executed recently and not high memory is consumed in Delta. If your server size is not adequate, you may look into removing some data from the tables.

Also, check if server restart is possible. (actually a dumb suggestion, but without knowing any further details, can't think of anything else. BTW, sometimes even the restart works)

Regards,

Ravi

[sunny_pahuja2]

Hello,

As per screen you shared, Disk has sufficient space and problem you are facing is not related to Disk space. You should check memory, as Database used memory is very high. It might be the case that you need to increase memory on the host or may be you can add another host. But prior to that you need to find out which process is taking so much memory. Please refer to SAP KBA 1840954 and it will guide you to find out if there is any problem in the HANA box or you might need to increase memory.

Thanks,

Sunny