Re: Journalling All Files

The SQL SCHEMA [or COLLECTION] journaling interface is activated simply by the existence of the object name QSQJRN as *JRN object in the library; i.e. creation by SQL is not a requirement. However only the SQL CREATE TABLE [database *FILE object] is affected by that type of automatic\implicit journaling implementation. Also be aware that having more than one *JRN object on the system [iASP really] with the same name is discouraged, since /recovery/ for the journal objects is impacted; an /oops/ in the original design :-( Ideally, only the one QSQJRN in the QSYS2 library of the iASP would exist on a system.

For a more generic library-based automated\implicit [object] journaling implementation, investigate the QDFTJRN *DTAARA which by its existence [similar to the QSQJRN *JRN] with its character data will identify the qualified journal name [i.e. not limited to within the same library as with the SQL implementation], the object type(s) affected, and some detail for the level of journaling. The feature is not part of the CRTLIB & CHGLIB, which IMO seems much more appropriate; probably best to have been implemented that way, optionally identifying an exit program by a special value *REGFAC versus trying to identifying all of the journaling details via parameters on the XXXLIB commands themselves.

http://www.google.com/#hl=en&q=qdftjrn+site%3Aibm.com
http://publib.boulder.ibm.com/infocenter/iseries/v6r1m0/topic/rzaki/rzakiautostrjrnl.htm
http://www.redbooks.ibm.com/abstracts/tips0604.html
http://www.redbooks.ibm.com/abstracts/tips0662.html

FWiW there is some limited reporting available with the CMPJRNIMG [compare journal immages] command. As for journaling temporary files, I think that is mostly an application decision. Journals are not nearly as friendly for temporaries as for persistent\permanent object for which the current\existing object identifies the current JID [journal identifier] against which image selection is done; e.g. a renamed file requires knowing timestamps that the file existed and searching across the separate names.

Regards, Chuck

Alan Campin wrote:

If you create a library as an SQL library any files that is added
to the library is automatically journaled. You can do this be
using STRSQL and Create Collection or by using Ops Nav to create
a schema.

On Sat, Sep 12, 2009 at 1:30 PM, Jim Franz wrote:

I've been asked to comment on the feasibility of a request to
"journal all files" for security and record change audits. This
is a system I'm very familiar with, and already has some
journaling and trigger based auditing. I've just never had a
management request to journal "everything". (New owners-and
prob in Fortune 100) I should add this is a very customized software package, call it a "mrp/job shop", 500-700 users, and
does answer to corp, government, and military auditors.

Anyone out there journalling "all", and what the best way to
provide a user interface to querying the receivers to (for
example) list who changed a certain field over the last year.

Do temporary work files count?