Subject: Re: G-ZIP test
From: Erotokritos Katsavounidis (kats@hep213.cithep.caltech.edu)
Date: Tue Jun 20 2000 - 19:29:07 EDT
Nicola,
It is not clear to me all the details of what you did. Here's some
issues that you might got wrong and are responsible for the problems
you are having:
1) the source GC file has to be RAW with VARIABLE record length,
exactly as it is stored in the VMS COPY of a typical NGC tape
(so this is wrong for GC011111.DST you worked on).
2) If you gzip'ed the GC file after changing its record length
to a fixed one, you got it wrong.
3) In general, you should never change the FILE ATTRIBUTES of
input/output files before/after you gzip/tar/untar them.
4) I believe I had mentioned in my e-mail yesterday that you should
refer to my DISK$MACROSCRA3:[MACROUSA0.BACKUP.MUDSTS]TAPEMAKER.COM
file that has succesfully transfered in the part the MUON DSTs
from disk to tape.
Did you do that? You should look in the COPIER subroutine
where you'll see that the steps you should follow are:
o gzip <onefile>
o vmstotar cv <onefile>-gz
i.e., VMSTOTAR *is* the COPY command that takes a file from
disk (the gzip output of file <onefile> called <onefile>-gz)
and puts it on the tape drive that your processed has allocated.
If no tape drive is allocate, VMSTOTAR creates a (Tape ARchive=TAR)
file on disk that it calls TAPE.
In order to unfold the TAR file you have to use TARTOVMS. You
can NOT unzip a TAR file. UNZIP command will NOT accept as input
files that were produced from the VMSTOTAR command.
I am rather disturbed by the time it takes to make progress on
rather simple and straightforward things. Please first thing in
the morning, use DISK$MACROSCRA3:[MACROUSA0.BACKUP.MUDSTS]TAPEMAKER.COM
to upload a RAW GC DST from DISK to TAPE on the VAX. Then, in order
to be 100% sure, download the file on CAMELOT's DLT unit (ask
Sandra Parlati to help if you don't know how to untar the tape
on unix). I would really like to see the disks empty from the RAW GC
DSTs by tomorrow.
--Erik
This archive was generated by hypermail 2a24 : Tue Jun 20 2000 - 19:29:14 EDT