Q: Are the server and job logs available in raw format? Can I access this data directly?
A: Yes you can. Server and job data is stored in the repository for the cluster. In the database browser you can point to the cluster being used and see all the tables for that cluster, including job information, server events, etc.
All data is available as raw data inside these tables and you can transfer it wherever you want. In fact, you can use a Centerprise dataflow to send it to any destination of your choice.
Q: Is the repository built as a part of the installation? Where is it stored?
A: First you need to build the cluster database for the repository. You can do that by going to the server menu and clicking on “Build Cluster Database.”
In this command, you will be required to point to an SQL server database, where the cluster database will be created.
Once the cluster database is created, you can use the connection information to configure a server and tie the server to the cluster in question.
Q: Can I shut down certain components of a server and does it make the server any faster?
A: The server profile of a server shows all the active components of the server. By default servers are tied to a ‘default’ profile. If you have changed the profile of a server, you can go to the server profile maintenance screen of a cluster and look at the details of the concerned server profile.
You can configure the components of the profile in the question by selecting or unselecting components from the active components list.
Does this make the server faster? It depends on the active components. Some of the components are heavy and some are lighter in terms of the resources they require to run. The job processor, for instance, is a heavy component, so shutting it down will free up a lot of resources on that server and make it faster. But if you are talking about a lighter component such as the failover processor, that kicks in only when another server fails in the cluster so because it is lightweight, switching it off most likely will not affect performance significantly. In short, server performance depends on which components you are planning on switching off.
Q: Can I export a filtered view of the job monitoring dashboard?
A: Yes. Looking at the job monitoring dashboard screenshot below, it is a filtered view showing only the successful jobs from this month on all the servers.
This view can be exported simply by selecting everything that has been filtered and using the context menu command to export to Excel.
That is how a specific view can be exported from the job monitoring dashboard, but if your requirements are more involved, you can export the raw data directly from the repository table by using a Centerprise dataflow.
Q: How do you rectify directory paths when load balancing occurs? The file in Server A may not be in Server B.
A: We recommend using UNC paths for your flow files when you schedule them so that they are accessible from each of the servers. For example, if Server A is scheduled to run a job and it goes down, the failover processor transfers this job to Server B. Server B should be able to access the location of the flow files for this job, and if it cannot access that path, it will complain about file path not being accessible.
Q: The job log table can fill up quickly. How do you recommend purging and archiving the data from these logs?
A: The default value for purge frequency is seven days. The cluster database purges all the log data based on this frequency to keep the log tables clean. If you choose to archive the log data, we recommend using a dataflow to move the log data to the archive before purging.
Q: Is there any notification for a situation when a server dies and its jobs are transferred to another server?
A: Yes. The server log contains this information. A log entry is made by the failover processor about the failed server and components of the failed server taken over by the new server.
Q: Can you explain how job priorities work?
A: In the job management screen, under server jobs, you can see what the priorities are for any running jobs.
Centerprise supports three priorities: high, medium, and low, and by default all jobs are assigned medium priority. If you want to change the priority of any of the queued jobs, you can use the context menu command to raise or lower the priority of the job.
0 Comments