Archive infostructure customizing when active

[jespergk]

Hi Experts

We have archived object ISU_BCONT using standard infostructure SAP_ISU_BCONT.

When we try to read from archive through transaction SARI we can see the list, but it is not possible to open object (business contact). By debugging I found out that the infostructure does not have any values in the two customizing options "Field name Processing" and "Offset Processing". It works fine in our test system which have a "D" in both ("File name is data field" and "Offset is data field"). The customizing table is AIND_STR1.

I am guessing it is not possible to change the infostructure after the archiving have been done? Even changing a standard infostructure might not be possible?

We have also tried to use the report "REARCH_BCONT_READ_SEQ". That seems to try to open all archiving files in the beginning, but that also fails which seems to be a simple timeout - as far as I can tell it loops through ALL records which is not feasible. This even when a specific object and date is given as input.

Any help would be deeply appreciated!

[3070856]

Hi Jesper.

It is possible to "change" a customer info structure after archiving has started. It is actually quite flexible.

There is a "copy" function that will even "move" (re-assign) the entries from the one reference to the other. (I have used in partitioning but don't see why it wouldn't work elsewhere. It has a test function you can use to check). You can also delete and rebuild the entries in the info structure if all data is intact.

You are correct that you should never change a SAP std info structure or catalog. Always create your own. (Catalogue only if required).

As this post is quite old, I'll leave at that for now.

[jespergk · Accepted solution]

So we are not sure if we will do this, but I expect the following procedure would fix it. All this is available through transaction SARI and it's menues.

1. Deactivate the infostructure SAP_ISU_BCONT

2. Delete data

3. Delete program

4. Activate Infostructure

5. Fill the Infostructure (this is client dependent and will fill table ZARIXIS1)

The reason why we are a bit hesitant is because of step 5. We have clients with close to 600.000 archived contacts, so filling the structure could prove to be very time consuming or might even dump.

We are considering creating a custom report based on standard REARCH_BCONT_READ_SEQ, as that actually seems to work without these offsets, but not when too many files are opened at once. We would then use the parameters to only open the relevant files (these can be found directly in table ZARIXIS1 which should be a quick SQL search). This of course would mean the customer would have to provide reasonable search parameters, but it is much less risky.

I will report back if we decide to refill the infostructures. Hope someone can benefit from this at some point!

[jespergk]

So we went ahead and did it, and thankfully it worked without issues and the offset was added as expected based on the customizing. First it would not activate the program but I think it was because we forgot to delete the tables.

Filling the structures went pretty fast - our biggest client had ~500k archived objects and that took about 15 minutes.

Now we can finally access the archived objects from SARI with good performance.

[3070856]

Ignore my previous responses - glad it came right

[jespergk]

Thinking out loud on text, so be warned 🙂

So it seems the infostructure has generated table ZARIXIS1 without the offset field "ARCHIVEOFS" because of the wrong costumizing mentioned above.

From SARI customizing I can delete the entries in the table and delete the program after inactivating the infostructure:

When this is done I might be able to change the customizing somehow, more on that later.Then I activate the structure and try to fill it again and this actully seems to fill table ZARIXIS1 again with correct offset (at least in our test system which already had an offset, so might not be the case in production):

That still leaves me with the question - how can I change the infostructure customizing? The table behind is client independant: AIND_STR1. When I try to customize it from our development system, those fields are not open for editing.

That led me to "Field Catalogues" and the table "AIND_SV1S" which seems to be what dictates these values. To my surprise it seems that this is setup correctly:

So how can it be that the infostructure is not generated correctly when the field catalogue specifies it? Maybe it is timing, would you dare to try in a production client? (inactivate, delete table, delete program, reactivate, fill structures)