iCluster Journal Analysis Report Part 3 – Sample Reports
In our last post, we listed the 8 spool files produced by the DMANZJRN command used for journal analysis. In this post we look inside the first 4 of those spool files.
DMANZJBAND – this is in essence a report showing the bandwidth used by journaling and replication of the objects that you already have journaled. Your actual bandwidth requirements may be different if you have objects or files that are not yet journaled. It lets you know where the peaks and valleys are during the period captured. It is best to run this command over a heavy batch period like quarter end or month end processing as well as again during a normal day. As you can see in this sample, it was a quiet period with only one peak on May 26th at 5:00am. It shows how many database changes, how many bytes per hour were generated and the % of the line based on various line sizes that would be utilized.
DMANZJNTVO – this report shows the information based on object, in other words what file or native object is being hit the heaviest in terms of number of transactions and the totals amount of data. In the sample below it is file WATERLEV in library GWDB at hour 17 on May 28th.
DMANZJBSFO – this is the equivalent of the DMANZJNTVO report but for BSF objects.
DMANZJPGM – this report tells you what program did the most damage. In this sample there are 2 programs, BKUPWEEKLY at hour 3 on May 24 and QCMD in hour 17 on May 28. The latter indicates someone from the command line is making the most changes.
In our next post we will cover the remaining 4 reports.