Fileset count mismatch

[Denis R. Dufault]

Hello,

I am preparing to patch the Prod server and have an issue matching the fileset count between the create_depot script, swinstall -p and swlist -l fileset commands.

I just built a depot from a recent patch assessment with PH patches and the GOLDQPK11i depot.

My concerns are:
- why does the result of swinstall -p have such low fileset numbers compared to create_depot and swlist cmd ? Does swinstall group the filesets into subproducts and t hus report a fewer filesets ?

The swjob cmd after a create_depot script gives:
* 1151 of 1151 filesets had no Errors or Warnings.

The swjob list after an swinstall -p -d command shows:
* Summary of Analysis Phase:
* 34 of 341 filesets will be Skipped.
* 307 of 341 filesets had no Errors or Warnings.

And, the swlist -l fileset gives:
swlist -l fileset -d @ |grep -v "#"|wc -l

1152

OR: 1583 - with dependencies (#) included.


Thanks,
Warren

[Dennis Handly · Accepted solution]

Either you or I'm confused. (I hate it when commands have two different ways of doing things.)

I found you never use -d with SD except some very few commands. Instead use -s for source and "@" for target.

So what do these say:
# swinstall -p -s
# swlist -l fileset -s |grep -v "#"|wc -l

Also swlist will list everything and swinstall will only list what can be installed or hasn't been installed.

[Denis R. Dufault]

Thanks for the information on what swlist and swinstall returns. I can live with the fileset counts now. Also, sorry I did not mention I have the patch_match_target=true, which is probably the reason for your confusion and mine. The numbers look better using the command:

swinstall -p -s \*

with result:
* Summary of Analysis Phase:
* 594 of 898 filesets will be Skipped.
* 304 of 898 filesets had no Errors or Warnings.


Yes two ways to do the same command is confusing to me. I now understand I can write the same command as:
swlist -l fileset -s |grep -v "#"| wc -l

swlist -l fileset -d @ /patches/april32007/depot|grep -v "#"|wc -l


thanks again