Reviewing configuration and computer settings

Resource limit requirements on AIX and Linux

When resource limit requirements such as fsize and nofiles are not configured correctly, a gradual system slowdown can occur, even before errors are generated. Ensure that the resource limits are configured correctly. See Configuring IBM Content Search Services servers.

Disk space

When the disk on which IBM Content Search Services is running begins to run out of space, overall system performance gradually degrades even before out-of-disk-space errors are observed. For example, you might experience a degradation in performance due to operating system temporary folder cleanup. Ensure that you have enough available disk space for the IBM Content Search Services temporary folder and for your collection data.

Disk and file system configuration

IBM Content Search Services works intensively with the disk by reading program resources, reading and writing configuration data, logs, indexed data, and temporary files. Even when disk I/O is not a bottleneck, using a faster disk for IBM Content Search Services can improve overall indexing and search speed for the program files folders, configuration and logs, collection indexes, and the temporary folder.

When considering disk speed, take into account connectivity to the disk device. For example, when IBM Content Search Services files are stored on a remote disk with high latency or limited throughput, indexing performance can degrade.

Disk block size can also affect performance. For example, in some environments, a disk block size of 64 KB provides optimal indexing throughput. While a large disk block size can improve indexing throughput, a smaller disk block size might improve the search performance. Ensure that you set the disk block size to an appropriate value for your environment.

When running searches against large collections, the search uses the disk cache. As a result, search response times can be improved by increasing the size of the system memory (RAM).

Disk configuration can be different for different file systems. For example, in some environments that use GPFS, you can improve performance by activating the maxFilesToCache parameter and increasing the value of the pagepool parameter to 4 GB.

IBM Content Search Services threads

A smaller number of indexing batches than configured leads to a noticeable degradation in the overall indexing throughput rate

Because obsolete Virtual Servers in the Global Configuration Database (GCD) are not automatically cleaned up, that Virtual Server count can be higher than the actual number of CE instances with dispatching enabled. That inflated number will then result in a smaller number of concurrent batches per Content Search Services server, negatively affecting indexing performance.

Besides degradation in overall indexing performance, you might also see a number for "Total CE instance configured for indexing" in the summary log that is higher than the number of CE instances that are enabled for indexing dispatching on the system. Additionally, you might see that the list of Virtual Servers in the Administration Console for Content Platform Engine is too high.

For more details, see Content Platform Engine uneven CBR indexing workload and indexing degradation .