🚀 go-pugleaf

The reports generated by Cheyenne's "ARCsolo for Windows" have entries
which are badly wrong or completely missing. Details follow...

Configurations
==============
  In various combinations:
         ARCsolo 3.01a and 3.02
         Various PCs
         SCSI Adapters: Adaptec 1542CF or Future Domain 1680
         DAT: HPC1533A (firmware revisions 9312 and 9406)


Flaw 1: "Blocks Written" fields wrong.
======================================

In the DEVICE and MEDIA reports, the following fields are erroneously
blank, zero or orders of magnitude too small:
  Blocks Written
  Total Blocks Written
  Usage
  Total Usage Time
 
After tens of e-mails and faxes with Cheyenne the problem was fixed by
sending me replacement DEVICE.RPT and MEDIA.RPT files. This was over a
month ago and yet they have not placed these files on their patches
archive. I wonder how many other fixes are not advertised. I will
happily mail the patch to anyone who asks.


Flaw 2: 
=======

Whether or not the above new report generating files are used, the
MEDIA ERRORS ENTRIES ARE NOT TO BE BELIEVED. During many tens of
gigabytes of backing up data on my system, ARCsolo has not reported a
single media error, corrected or otherwise. This, despite finding a
very small percentage of files are corrupt on the tapes.

Cheyenne's contradictory quotes (chronological): 
  Chey: "As for errors, you should see some for no tape is made perfect."

  Me: "Does the HPC1533A, ... report media errors correctly to ARCsolo?"
  Chey: "Yes, it is suppose to when working properly."

  Me: "Is it normal to have had NO media errors at all (recovered or
       otherwise)?" (over tens of gigabytes of backing up)
  Chey: "Yes."

  Me: "Does ARCsolo really report media errors from an HPC1533A?"
  Chey: "No, the report file only reports at random specific times during 
         the backup.  Unless you get media errors on every file that is 
         written you will probably never see one being reported."
  [What on earth is a "random specific" time?]

Even accepting their later statements that media errors are not
reliably reported [Has anyone EVER seen ONE reported?], the claimed
"detailed reports" are at best misleading.


I am now resigned to the belief that Cheyenne apparently have no
intention of providing sensible media error information.
     

----------------------------------------------------------------------
Pete B. Wilson                            Signal Processing Laboratory
E-mail: pbw@eng.cam.ac.uk                 Engineering Department
Tel:    +44(1223)332767                   Cambridge University
Fax:    +44(1223)332662                   Trumpington Street
http://www2.eng.cam.ac.uk/~pbw            CAMBRIDGE ENGLAND CB2 1PZ