Throttling and multiprocessing in Secondary

To keep the data in the Primary and Secondary server up-to-date, the Primary writes to its transaction log and forwards it to the Secondary so that the Secondary can make the same changes to its copy of the database. If the Secondary cannot keep up with the processing pace of the Primary, the solidDB^® throttling mechanism slows down the processing on the Primary. From the application standpoint, throttling results in increased response times.

As of V7.0, the Secondary can use multiple threads for processing the write loads. The Secondary receives operations from the Primary in one stream. Operations are parsed and each transaction is attached in execution queue; there are as many execution queues as there are open parallel transactions. To ensure that conflicting operations are executed in the correct order, row locks are used.

Perfmon variable	Description
HSB grpcommits	Number of transactions in the most recent group commit Transaction commits are grouped in one-log bursts which are sent to Secondary as one packet. You can use this counter only on the Primary.
HSB secondary ops in packet	Number of log records the Secondary received from the Primary in the most recent log record packet.
HSB secondary trx count	Number of open transactions the Secondary has received from the Primary
HSB secondary locks	Number of row-level locks on the Secondary
HSB secondary lock reqs	Number of lock requests on the Secondary
HSB secondary lock waits	Number of lock waits on the Secondary since the server was started
HSB secondary op waits	Number of times operations (transactions) on the Secondary have been waiting to continue execution
HSB secondary buffers	Number of buffered log record packets the Secondary has received from the Primary
HSB secondary serial mode count	Number of times the Secondary parallel executor has switched to serial mode instead of running in parallel
HSB secondary dispatch queuelength	Size of the most recent dispatch thread (operations to dispatch) on the Secondary

You can also optimize the multiprocessing level of the Secondary with the HotStandby.SecondaryThreads parameter. The optimal number of threads depends on the environment and requires experimenting with loads typical to your environment. In principle, the more processing capacity (cores) the computer has, the higher you can set the value of the HotStandby.SecondaryThreads parameter. However, excessive use of threads is not likely to improve performance.