When Qlik Gold Client import jobs seem to be running long, here are a few things to check for possible areas of improvement
- Check how many background processes are available in the target system and how many application servers are installed. If there are a very low number of background processes configured, Basis can investigate adding a few additional so the import jobs can be spread over more jobs. In conjunction, check the Parallel Processing configuration in Gold Client found under ZGOLD – Configuration – Administration. This setting determines how many background processes per application server can be used for Gold Client jobs. There should be a good balance of how many are available and what is configured. You want to be careful not to consume all background processes, however, in situations where you are building a new client and no one else is using it, this number can be set higher until the new client is ready.
- Is a data replication process such as Qlik Replicate or SAP SLT running in the target system? If so, it is best practice to suspend this while large loads are importing to avoid resource contention.
- If this is a client refresh and no other users are logged in and transacting, archive logging can be disabled until after the data loads are complete. This also helps with performance since there is no risk of consuming all space in archive log folder as well as avoiding resource contention.
- If the target system is a Sandbox or another system with light resources, it is possible Basis may need to investigate adding additional resources or parameter changes if it is a heavy data load.
- Gold Client’s Sizing Report can be in utilized for both Client Construct and Data Echo to gauge which data is the heaviest. You may find that you do not require all data for a given area and potentially skip it in future data loads. For example, when copying data with Client Construct, you may find a small number of tables that are comprising the majority of the data. In that case, exporting a slice of that data with Data Echo may be acceptable.
If you need assistance with data import performance, please open a support case (www.qlik.com/community-support-cases).
