Read process takes too much CPU resources. (4329709)

Monitoring of CPU usage by various processes on a server using "top" or any applicable command show that Read is taking too much CPU resources.

Increased CPU usage by Read process can be caused by either the increased activity on source database or very large *tcache* files in the source's $SP_SYS_VARDIR/state directory.

1) The problem can be due to either the increased activity on source database or very large sized *tcache* files in the source's  $SP_SYS_VARDIR/state directory. While there is nothing that can be done in case of former, in case of latter the following can be done:

a). Stop Read

b). Go to the source's $SP_SYS_VARDIR/state directory, issue "ls -l" (use equivalent command if on Windows or NT) to see if the files *tcache.a          and *tcache.b are very large, say of the order of tens of megabyte. If so, then rename them to something else so that when Read is restarted,          it will create new *tcache* files which will be much smaller.  NOTE: The first open transaction may be lost and this can result in some out of              sync on the affected table.

c). Start Read.

The above steps should result in Read taking less CPU resources.

WARNING:   If one cannot afford any out of sync on target tables, then the procedure is not advisable and one may have to let all the messages in the queue post first before trying the above measure. In that event, the problem will be prevented from occuring only in future.