Hello!
It would appear this particular fixlet (IBM BigFix - Updated Platform Server Components version 9.5.2 Now Available!) doesn’t use prefetch blocks to prefetch (it just uses normal prefetch commands) and so it always caches every possible file in the fixlet regardless of what is actually required.
Just wanted to drop a note…
Related to this: BigFix 9.5 Patch 2 is now available
I feel like incorrect might be a bit strong, but definitely less than ideal. I think this is a general issue in most of the IBM patch content that has multiple download options. Most of them should be using prefetch blocks instead.
I have had this complaint about the patching content in general for a while.
In order to leverage prefetch blocks in a way that would prevent caching every file on the Root Server associated with the Fixlet, we’d have to leverage dynamic downloading (which requires that the download whitelist be updated accordingly). It’s something we’re still exploring, but it is unfortunately not as simple as simply leveraging prefetch blocks.
Good to know! Thanks!
It was my (mis)understanding that dynamic downloading wouldn’t be required. Relevance could be used within the prefetch block to determine which downloads were required and which ones were not.
Example: https://bigfix.me/cdb/fixlet/3967
That is a bummer. Dynamic downloading wouldn’t be a great option in my opinion.
I guess I was mistaken in how I thought the IF/THEN statements in a prefetch block affected what was downloaded.
I suppose the solution I’m looking for is an option to not have the root prefetch the files until requested for the first time.
The IF/THEN blocks optimize downloads so individual clients only download what they need, but the server still prepares for any download request unless dynamic downloads are used. A somewhat simple optimization to try and prevent delays in action execution from external networks/hosts which are generally out of our control.