1 of 1 people found this helpful
Steffen, you've pointed correct folder where the Catalog metadata is located; every Catalog has <filserver>/patch/catalog/catalog_xxxxx folder with hf7b and pd5 xml files.
>>We have noticed that when we update the Depot items (hf7b.xml and pd5.xml) that are linked to the Patch-Catalogs, they still use the "previous" versions of the XML.
I did not understand this completely, but will still try to answer..
The Catalog will use the metadata that is available to it. If you made the 'previous' version available, then that's what the Catalog will use. I'll explain 'previous' later..
It sounded like you may be using offline catalog and if so, then whatever versions of hf7b and pd5 you downloaded and added to Depot, that's what will be used.
But if you use online catalog, then the Catalog will download the files using the URLs in patch global configuration / shavlik url configuration.
About downloading 'previous' version. Some infrastructures are behind web accelerators, which cache information from the Internet. If you are behind one of these devices, then it's very likely that you are downloading cached older copy of shavlik files, rather than newest version. You can use this test to validate at the time when the new shavlik xml was released:
wget --no-cache link/to/shavlik.xml
It's possible that you will download two different files, and the one with no-cache flag would be the newer one. More information about this here: https://kb.bmc.com/infocenter/index?page=content&id=KA323772
Now on to your main question, if it's possible to update the metadata without updating the Catalog. Technically yes it's possible, all you have to do it replace the files in the fileserver/patch/catalog/catalog_xxx folder. But this will not always work flawlessly, because if there are new patches added to metadata, then what you have right now is missing objects from the Catalog, and they will not be reported missing in Patch Analysis Jobs. I would still recommend to allocate the time to update the Catalog to bring all information in sync. I have personally done this before, don't recall seeing issues, but again, I was doing this in my lab where I constantly break things, rather than production environment.
thx for your feedback.
We are currently testing this in DEV and to see how that goes.