Re: SOX monitoring and journal interpreter

I have to agree with what several have said about using Turnover or a
similar tool for program changes.

I would have limits on programmer authority on the production box, and
perhaps only a limited set of programmers have any access, the rest don't
even have a profile or if they do it is a user profile.

I would disable DFU and utilize a tool like DBU that can have auditing of
changes made within the tool. In some manner, a tool that would the staff
would log their changes. Name, Date, Time, What was changed, why was it
changed, who requested the change would be the basic elements that I would
track. Compare the log to the audit file of changes on a regular basis
(depending on your situation this could be daily, weekly, bi-weekly,
monthly, etc.) Audit shows a change not logged would require disciplinary
action. Track the monitoring to satisfy the auditors.

After my first two years, I have pretty much been the lone programmer,
admin, operator, etc. on a single box/single partition. I have worked in
some very highly regulated (SOX, USDA, etc.) and non-regulated environments,
so I have learned to be careful about making production changes. I have
modified code to take advantages of the system to minimize documentation
requirements (changing hardcoded dates in a program that have to be changed
every year, to utilize a data area as an example-changing the value in a
data area is much less paper work than changing a hardcoded date in a
program).


subject: [WDSCI-L] SOX monitoring and journal interpreter

Thanks to SOX we're asked to develop a system to monitor activity by
developers on the production system (An IBM i). First, has anyone else done
something like this, and in an overview, how did you do it?

For my attempt, I'm summarizing file journal entries based on some criteria,
like user, and generating a summary report/spreadsheet with a concatenated
key. Then from that, I want to allow an auditor to drill into that summary
to get the detail of file changes. For now, I'm copy/pasting this "key" into
another report program that is going to report change by change, what
happened in that file for that user, job, etc.

Anyway, I'm using the DISPLAY_JOURNAL SQL table function, and I have this
huge and varying format ENTRY_DATA column. This is where the actual file
data is. I'd like to split this into the file's columns on the spreadsheet,
which I could do if it was all character data, but I'm having trouble
figuring out how to split numerics of varying size and format into something
in an RPG variable. If I use %subst to break a chunk of the buffer out, what
kind of field would I feed that into, because the datatypes will vary so
much?


___________________________________
Darren Strong
Dekko