Nonstop Database, HiRDB Version 9 System Operation Guide
This section explains the items you need to revise when the number of users increases.
When the number of users increases, revise the values specified for the following operands:
The size of the shared memory used by HiRDB increases, which might result in shared memory allocation errors. In such a case, the shmget() system call error is reported with the KFPO00113-E message. Appropriate action must be taken on the basis of this message.
For details about how to determine the values for the OS system parameters, see the HiRDB Version 9 Installation and Design Guide.
For details about how to modify the OS system parameters, see the applicable OS documentation. When a system parameter is modified, the new value does not take effect until the OS is rebooted.
For information on determining values for shared memory and semaphores, see the HiRDB Version 9 Installation and Design Guide.
As the number of users increases, the number of ports used by HiRDB also increases. If the number of ports becomes inadequate, processing might be interrupted or the communications processing of other programs might be affected adversely. The number of ports assigned automatically by the operating system depends of the OS, so you should check the network-related settings on each server machine. For details about the number of ports that HiRDB uses, see the HiRDB Version 9 Installation and Design Guide.
If the number of ports assigned automatically by the OS is inadequate, use the pd_registered_port operand to specify a range of port numbers to be used by HiRDB. For details, see 9.7 Specifying a range of port numbers for use in communication processing.
When the number of users increases, the number of locking requests also increases. As a result, an insufficient locking pool size might cause an error (the KFPS00443-I message is output). In this case, check and, if necessary, revise the value specified for the pd_lck_pool_size operand.
If a locking pool size shortage occurs at the front-end server, check and, if necessary, revise the value specified for the pd_fes_lck_pool_size operand. The KFPS00443-I message shows the server where the locking pool size shortage occurred.
In the case of a HiRDB single server configuration, more semaphores and semaphore identifiers are used, which means that the semget() system call error is reported with the KFPS01815-E or KFPO00107-E message. Appropriate action must be taken on the basis of the message:
In the case of a HiRDB parallel server configuration, an increased number of users has no effect on the number of semaphores or semaphore identifiers.
For details about how to estimate the memory used by the OS system parameters, see the HiRDB Version 9 Installation and Design Guide.
The synchronization point dump file and status file sizes increase. For information on determining the sizes of synchronization point dump files and status files, see the HiRDB Version 9 Installation and Design Guide.
After determining their sizes, re-create synchronization point dump files and status files.
If the number of users increases, determine the revised values for the OS system parameters. For details about how to determine the values for the OS system parameters, see the HiRDB Version 9 Installation and Design Guide.
In the AIX edition, you need to also increase the values for the following OS system parameters:
If the number of users that were authorized at the time HiRDB was purchased is exceeded, HiRDB must be upgraded.
All Rights Reserved. Copyright (C) 2011, 2015, Hitachi, Ltd.