SUMMARY: Fverify question

From: Ronald D. Bowman <rdbowma_at_tsi.ctd.ornl.gov>
Date: Sun, 15 Feb 1998 11:19:45 -0500

Hello managers -
        
        I would like to thank Dr. Tom Blinn for his quick response
        to my original post shown at the bottom. I guess I am not
        the only one working over the weekend. He confirmed what I
        thought was the situation with fverify. Running fverify on
        a software subset inventory that has been patched will result
        in checksum and file size errors. He informed me that DEC
        knows about this problem and that it "results from the way that
        the patch installation tools manage the software subset
        inventory records". Or actually, I guess the patch tools do not
        make a merged inventory file for those software subsets that
        are patched. We can hope that future releases of fverify will
        take this into account.
        
        So, if you run fverify on a system that has been patched, you
        can expect to see checksum and file size errors. Do not
        initially conclude that something is wrong(something I did
        initially).

Sincerely,
Ron Bowman
Techno-Sciences, Inc.
864-646-4028
864-646-4001(fax)
rdbowma_at_tsi.clemson.edu

Original Post:
===================================================

Hello Managers -
        
        Sorry about the length. The question is marked a few paragraphs down.
        
        I have a question concerning fverify. After installing Pathch #6,
        I decided to run fverify on all the inventories in /usr/.smdb./.
        What I found was the following(after a little extra work) -
        All patch files verified ok. All software subsets that were not
        patched verified ok. However, all software subsets that were
        patched had checksum and file size errors.
        
        
        Further investigative work on this last condition, led us to the following
        conclusion: When patches are installed various files in a software
        subset are modified(thus changing the checksum and file size), the
        inventory list for the software subset is not changed(It can't be
        in case the patch is removed). Therefore to correctly check a particular
        file in a software subset you would have to know if a patch modified it, and
        verify that patches inventory. Plus you would have to manually confirm
        files that failed in the software subset inventory verification were
        successfully verified in the patch inventory verification.
        
        From the man page on fverify there is a way you can mark particular files,
        and then use a flag with fverify to ignore changes to those files. That
        sounds great if it is only a few files. However, that is more tedious if
        it is a large number of files modified by several patches.
        
        QUESTION:
        Is there a way to use fverify on a software subset
        such that it also takes into account patches that have been applied?
        I would like to be able to verify a software subset such that fverify
        would determine if a patch has been applied. Then if a patch has been
        applied, the files modified by the patch would be verified using
        the patch inventory entry instead of the software subset inventory entry.
        
        If it is not possible to currently verify patch modified software subsets, it
        would be nice if that feature could be added in future versions of fverify.
        
        In my case too little knowledge on something led me to think that something
        was wrong when in fact everything was normal.
        
        Any explanation about how to get(or not get) fverify to work the way
        I would like is greatly apprecitated
        
        
        
Received on Sun Feb 15 1998 - 18:56:40 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:37 NZDT