Resource considerations for a partitioned Db2 server

When running NSE in a partitioned environment, you should keep the following resource considerations in mind:

• When multiple text indexes are updated in parallel, this can lead to considerable burst disk usage during the I/O intensive phases of the index update. This requires careful consideration during the setup of partitions, the associated NSE index storage, and scheduling of index updates.
• For NSE administration commands, there are processes started for every partition to carry out operations such as index create, drop, and update. The update process can be a longer running process that will consume resources. For scheduled updates of text indexes in the system, several index updates running concurrently can lead to as many processes per partition as the number of indexes being updated at a given time. It would be advisable to minimize the number of concurrent index updates by planning the index update schedule accordingly.
• On Linux® or UNIX operating systems, you can use the command ulimit with the appropriate options to view or change the size of the process resource limit. If you are running Db2 Net Search Extender in a partitioned database environment, use the command db2_all ulimit with the specific options to see the size that applies to all the database partitions. It is essential to verify the hard operating system ulimits for successfully running Net Search Extender update index commands. For example, insufficient data segment size on any of the partitions can cause failures during index update commands most usually with CTE0105 error logged in the event view of the text index.