eG Help

Although SAP HANA is an in-memory database, I/O still plays a critical role for the performance of the system. From an end user perspective, an application or the system as a whole runs slowly, is unresponsive or can even seem to hang if there are issues with I/O performance.

If a datafile is able to process I/O requests to it quickly, it is a sign of the good health of the SAP HANA Database Server. On the other hand, any slowdown in IOPS could indicate a serious processing bottleneck on the server. This can be due to many reasons for instance, synchronous IO behavior of the system. That means, triggering an IO request takes just as long as executing it and hence very prone to performance degradation of HANA system. In fact, if the I/O process is not properly tuned this could slow your SAP HANA system. That means right from restarts, to table loads to savepoints, delta merges etc. operations will be slowed drastically. When all these mission critical operations start lagging it will eventually cause SAP HANA system to slow down. This means that SAP HANA is having synchronous behavior, which is unacceptable. Administrators should hence continuously track the reads/writes to every files on the SAP HANA Database Server, identify the rate at which data related to these files was read/written from the buffer, the rate of synchronous reads/writes, and measure how well the requests were processed by each files. For this purpose, you can run the HANA IO Test.

This test auto-discovers the files types on the SAP HANA Database Server and reports the data read and write rate and rate at which reads/writes were performed on the files, and rate of synchronous read and writes. This test also reports the number of failed reads and writes, read and write latency, and number of blocked write requests. In the process, I/O processing bottlenecks can be detected and the files that are affected can be identified.

Outputs of the test: One set of results for each file type in the target database server being monitored

Measurement	Description	Measurement Unit	Interpretation
ReadSize	Indicates the rate at which data was read from this file type during the last measurement period.	GB/Sec	Compare the value of this measure to identify the contained files that is the slowest in responding to read requests.
WriteSize	Indicates the rate at which data was written to this file type during the last measurement period.	GB/Sec	Compare the value of this measure to identify the contained files that is the slowest in responding to write requests.
TotalIO	Indicates the rate at which read and write operations were performed on this file type during the last measurement period.	GB/Sec
ReadRate	Indicates the rate of synchronous reads from this file type during the last measurement period.	Reads/Sec
WriteRate	Indicates the rate at which synchronous writes occurred on this file type during the last measurement period.	Writes/Sec
ReadLatency	Indicates the total time taken to read from this file type during the last measurement period.	Seconds/read	A high value for this measure could indicate a bottleneck while reading from the files.
WriteLatency	Indicates the total time taken to write to this file type during the last measurement period.	Seconds/write	A high value for this measure could indicate a bottleneck while writing to the files. By comparing the value of this measure across contained filess, you can identify the data file to which write operations are taking too long to complete.
BlockedWriteReq	Indicates the number of write requests to this file type that were blocked during the last measurement period.	Number
FailedReads	Indicates the number of read requests that this file type failed to process during the last measurement period.	Number	A low value is desired for this measure.
FailedWrites	Indicates the number of write requests that this file type failed to process during the last measurement period.	Number	A low value is desired for this measure.