[pca] PCA Reports

[little help]

Martin,
   Agreed.  However, as you know management loves charts and percentage
graphs.  I just need some type of measurable indicator as to how much/often
we are patching hosts in the environment.  The option you provide below
seems to be valid but would also become a administrative issue as new
builds/images would skew the numbers all together.  I am thinking about
going through the xref file and trying to parse all possible patches for a
specific version - ie sparc Solaris 8.  This seems very ugly but unless I
can figure out another way this may be my only option.

On Fri, May 20, 2011 at 3:27 AM, Martin Paul <martin at par.univie.ac.at>wrote:

> little help wrote:
>
>>    I am running some reports via PCA and had a quick question about the
>> "List:" portion of the output that s generated via PCA. I am trying to
>> generate a report of missing patches against all possible patches for a
>> host.  For example can I get pca to report that I am missing 20 out of a
>> possible 200 patches for this host.  I need to report what percentage of
>> patches are missing from a host.
>>
>
> Determining the number of missing patches is simple - just look at the
> first number X after "List: missing (X/Y)". The number Y is actually kind of
> useless; I seem to remember that it was proposed by some PCA user as a rough
> indicator of how outdated a system is.
>
> As for the number of possible patches for a system, it's not so easy to
> find a concrete number for that. Using PCA's --fromfiles option, you could
> feed an empty "showrev -p" output to PCA and then look at the number of
> missing patches again. This is only half correct, though, as any Solaris
> update version (like Solaris 10 9/10) already comes with a number of patches
> pre-installed. So actually you would need to save "showrev -p" output right
> after installing the system, before any patch gets installed, and use that
> as input to PCA.
>
> I'm not sure whether this percentage of missing vs. applicable patches is
> really a useful number anyway. After all the only interesting number is how
> many patches are missing. And maybe not even that, as the importance of
> impact of two patches can be very different - a system missing a kernel
> patch might be assumed to be more outdated than one which is missing two
> patches for some obscure packages which no one uses.
>
> Martin.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.univie.ac.at/pipermail/pca/attachments/20110520/ac5d5fc5/attachment.html>