CC-tea/org-hyperion-cules
1
0
Fork 0
mirror of https://github.com/SDL-Hercules-390/hyperion.git synced 2026-05-02 00:24:23 +02:00
Code Issues Packages Projects Releases Wiki Activity
Files
3a4e9d920a786d00a45d3ebdae8a9bdb2b2785d1
org-hyperion-cules/html/hercload.html
Roger Bowler b6b3981e83 Correct outdated config statement in ibcdasdi example 
git-svn-id: file:///home/jj/hercules.svn/trunk@1389 956126f8-22a0-4046-8f4a-272fa8102e63
2003-05-30 12:23:43 +00:00

573 lines
20 KiB
HTML
Raw Blame History

<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 3.0//EN" "html.dtd">
<HTML>
<HEAD><TITLE>
Hercules Version 2: Creating DASD</TITLE>
<LINK REL=STYLESHEET TYPE="text/css" HREF="hercules.css">
</HEAD>
<BODY BGCOLOR="#ffffcc" TEXT="#000000" LINK="#0000A0"
VLINK="#008040" ALINK="#000000">
<h1>Creating Hercules DASD</h1>
This page describes various ways of creating and loading DASD
volumes for use with Hercules.
<h2>Using pre-built DASD images</h2>
<p>
If you can obtain one of the IBM Application Development CDs
which are supplied for use with the P/390, you will find that
these contain pre-built DASD image files which simply need to be
unzipped onto your hard drive. These images can be directly
read by Hercules. However, you cannot legally use this method because
the software on the CD is licensed for use only on the P/390
system that it was shipped with.
<h2>Restoring DASD from full-volume dumps</h2>
<h4>Creating an empty DASD volume</h4>
<p>
The dasdinit program must first be run from the Unix shell prompt to
create a file containing an empty DASD volume.
<p>
The format of the dasdinit command is:
<pre>
dasdinit [-options] filename devtype[-model] volser [size]
where:
options:
-z build compressed dasd image file using zlib
-bz2 build compressed dasd image file using bzip2
-0 build compressed dasd image file with no compression
-lfs build large uncompressed ckd file (>2G)
-a build dasd image file that includes alternate cylinders
(option ignored if size is manually specified)
filename name of dasd image file to be created
devtype CKD: 2311, 2314, 3330, 3340, 3350, 3375, 3380, 3390, 9345
FBA: 3310, 3370, 9332, 9335, 9336
model device model (implies size) (opt)
volser volume serial number (1-6 characters)
size number of CKD cylinders or 512-byte FBA sectors
(required if model not specified else optional)
</pre>
<p><br>
<a NAME="models">
The current list of device types and models supported is:
<p>
<pre>
CKD DEVICES
alt
devtype-model cyls cyls
2311 [*]
2311-1 200 2
2314 [*]
2314-1 200 3
3330 [*]
3330-1 404 7
3330-2 808 7
3330-11 808 7
3340 [*]
3340-1 348 1
3340-35 348 1
3340-2 696 2
3340-70 696 2
3350 [*]
3350-1 555 5
3375 [*]
3375-1 959 1
3380 [*]
3380-1 885 1
3380-A 885 1
3380-B 885 1
3380-D 885 1
3380-J 885 1
3380-2 1770 2
3380-E 1770 2
3380-3 2655 3
3380-K 2655 3
EMC3380K+ 3339 3
EMC3380K++ 3993 3
3390 [*]
3390-1 1113 1
3390-2 2226 1
3390-3 3339 1
3390-9 10017 3
9345 [*]
9345-1 1440 0
9345-2 2156 0
FBA DEVICES
devtype-model blocks
3310 [*]
3310-1 125664
3370 [*]
3370-Al 558000
3370-B1 558000
3370-A2 712752
3370-B2 712752
9332 [*]
9332-200 360036
9332-400 360036
9332-600 554800
9335 [*]
9335-1 804714
9336 [*]
9336-10 920115
9336-20 1672881
9336-25 1672881
</pre>
<br>
[<b>*</b>] size may be specified else size defaults to the
first listed model.
<p>
<h5>Volumes exceeding 2GB</h5>
<p>
For CKD volumes which exceed 2GB, such as the 3390-3, and the
<b>-lfs</b> parameter is not specified, the DASDINIT
program will create multiple files by appending the characters
<b><i>_1</b></i>, <b>_2</b></i>, <b>_3</b></i> etc.
to the file name specified on the command line. These characters
are inserted before the first dot (<b>.</b>) after the last slash
(<b>/</b>). If there is no dot, then the characters are appended
to the end of the name. Each file contains a whole number of cylinders.
Hercules CKD support recognizes the files as belonging to a single
logical volume. Specify the full name of just the first file in the
Hercules configuration file (e.g. "<i>filename</i>_1").
<p>
The DASDINIT program cannot create FBA volumes exceeding 2GB
unless the <b>-lfs</b> parameter is specified and <i>large file size</i>
is supported on your platform..
<h5>Examples</h5>
<p>
To create a 3330 model 1 CKD volume consisting of 404 cylinders
(plus 7 alternate cylinders too)
with volume serial number WORK01 in a file called
<b><i>work01.151</b></i>:
<pre>
dasdinit -a work01.151 3330-1 work01
</pre>
<p>
To create a compressed 3350 CKD volume consisting of 560 cylinders
(555 cylinders plus the 5 alternate cylinders)
with volume serial number SYSRES in a file called
<b><i>dosvs34.24f</b></i>:
<pre>
dasdinit -a -bz2 dosvs34.24f 3350-1 sysres
</pre>
<p>
To create a 3370 FBA volume with only 100000 sectors
(instead of the usual 558000 sectors)
with volume serial number WORK02 in a file called
<b><i>mini.work02.140</b></i>:
<pre>
dasdinit mini.work02.140 3370 work02 100000
</pre>
<p>
To create a 3390 model 3 (triple density) CKD volume
of 3339 cylinders with volume serial number WORK03:
<pre>
dasdinit triple.a88 3390-3 work03
</pre>
Because this volume exceeds 2GB, DASDINIT will create two files
with <b><i>triple_1.a88</b></i> containing cylinders 0-2518 and
<b><i>triple_2.a88</b></i> containing cylinders 2519-3339. If you
specify
<pre>
dasdinit -lfs triple.a88 3390-3 work03
</pre>
then DASDINIT will create a single file <b><i>triple.a88</i></b>
containing all the cylinders. Your platform must support <i>large
file size</i> to specify the <b>-lfs</b> option.
<h4>Formatting the empty DASD volume</h4>
<p>
After creating a DASD volume you can format it with a program
such as standalone IBCDASDI or ICKDSF.
<p>
Here is an example of the IBCDASDI control statements required
to initialize a 3330 volume:
<pre><code>
WORK01 JOB 'INITIALIZE 3330 WORK VOLUME'
MSG TODEV=1052,TOADDR=009
DADEF TODEV=3330,TOADDR=151,IPL=NO,VOLID=WORK01,BYPASS=YES
VLD NEWVOLID=WORK01,OWNERID=HERCULES
VTOCD STRTADR=1,EXTENT=5
END
</code></pre>
<p>
To run IBCDASDI, place the above statements in a file called
<b><i>init3330.txt</i></b> and start Hercules in S/370 mode with
a configuration file containing these statements:
<pre><code>
CPUSERIAL 001234
CPUMODEL 3145
MAINSIZE 2
CNSLPORT 1052
ARCHMODE S/370
0009 1052
000A 1442 ibcdasdi.rdr
000C 1442 init3330.txt
0151 3330 work01.151
</code></pre>
After IPLing from card reader device 00A, connect a telnet client
to port 1052, and press enter. At the IBCDASDI prompt, enter:
<pre><code>input=1442 00c</code></pre>
<h4>Loading the new DASD volume</h4>
<p>
Next you need to create a full volume dump file on your mainframe
and convert it to AWSTAPE format using the <b><i>tapeconv.jcl</i></b>
job in the Hercules source directory. The AWSTAPE file can then
be downloaded in binary format to your PC where it can be defined
as a virtual tape drive in the Hercules configuration file.
<p>
A standalone program can now be IPLed into Hercules to restore the
volume image from the virtual tape onto the formatted virtual DASD
volume.
<h2>Building a DASD volume from unloaded PDS files</h2>
<p>
The dasdload program can be run from the Unix shell prompt to
create a new DASD image file and load it with data from unloaded
PDS files.
<p>
The format of the dasdload command is:
<p><code>dasdload <i>[options]</i> <em>ctlfile outfile msglevel</em></code>
<p>
where
<dl>
<dt><em>[options]</em>
<dt><code><i>-z</i></code>
<dd>build compressed dasd image file using zlib
<dt><code><i>-bz2</i></code>
<dd>build compressed dasd image file using bzip2
<dt><code><i>-0</i></code>
<dd>build compressed dasd image file with no compression
<dt><code><i>-lfs</i></code>
<dd>build a large dasd image file (can exceed 2G)
<dt><code><i>-a</i></code>
<dd>build dasd image file that includes alternate cylinders
<dt><code><em>ctlname</em></code>
<dd>is the name of the control file which specifies the datasets
that are to be loaded onto the newly-created volume
<dt><code><em>outfile</em></code>
<dd>is the name of the DASD image file to be created
<dt><code><em>msglevel</em></code>
<dd>is a number from 0 to 5 which controls the level of detail
of the messages issued during the load.
</dl>
<h4>Control file</h4>
<p>
The control file required by the dasdload program is an ASCII text
file consisting of a <em>volume statement</em> followed by one
<em>dataset statement</em> for each dataset to be created.
<p>
The format of the volume statement is:
<p><code><em>volser devtype</em>[-<em>model</em>] [<em>cyls</em> [<em>ipltext</em>] ]</code>
<p>where:
<dl>
<dt><code><em>volser</em></code>
<dd>is the volume serial number for the newly-created volume
<dt><code><em>devtype</em></code>
<dd>is the emulated device type (2311, 2314, 3330, 3340, 3350, 3375,
3380, or 3390) for the new volume. FBA device types are not supported
by the dasdload program. Model may be specified like
<a href="#models">dasdinit</a> above.
<dt><code><em>cyls</em></code>
<dd>is the size of the new volume in cylinders.
If <code><em>cyls</em></code> is coded as <code>*</code>
or as <code>0</code> or is omitted, then the default size
for the device type and model is used. Cylinders is ignored
for compressed devices.
<dt><code><em>ipltext</em></code>
<dd>is an optional parameter specifying the name of a file
containing the IPL text which will be written to the volume.
The file must be in the form of an object deck containing
fixed length 80-byte EBCDIC records in the same format as
expected by IBCDASDI or ICKDSF.
</dl>
<p>
The format of a dataset statement is:
<p><code><em>dsname method units pri sec dir dsorg recfm
lrecl blksize keylen</em></code>
<p>where:
<dl>
<dt><code><em>dsname</em></code>
<dd>is the dataset name
<dt><code><em>method</em></code>
<dd>is the dataset loading method which can be one of the following:
<dl>
<dt><code>XMIT <em>filename</em></code>
<dd>the dataset is loaded from an unloaded PDS created by
the TSO XMIT command
<dt><code>SEQ <em>filename</em></code>
<dd>the sequential dataset is loaded from a <b>binary</b> file.
ascii/ebcdic translation is not currently supported. Also,
the dsorg must either be <em>PS</em> or <em>DA</em> and
recfm must either be <em>F</em> or <em>FB</em>.
<dt><code>EMPTY</code>
<dd>the dataset is initialized with an end of file record
(if DSORG is PS) or an empty PDS directory (if DSORG is PO)
<dt><code>DIP</code>
<dd>the dataset is initialized with a LOGREC header record
<dt><code>CVOL</code>
<dd>the dataset is initialized as an OS SYSCTLG containing
the minimum entries needed to IPL an OS/360 system
<dt><code>VTOC</code>
<dd>specifies the size and location of the VTOC. A dataset
name must be coded on this statement, although it is not
used. If no VTOC statement is present, the VTOC will be
placed after the last dataset on the volume and the size
of the VTOC will be the minimum number of tracks necessary.
</dl>
<dt><code><em>units</em></code>
<dd>is the space allocation units: <code>TRK</code> or <code>CYL</code>.
<dt><code><em>pri</em></code>
<dd>is the space allocation primary quantity
<dt><code><em>sec</em></code>
<dd>is the space allocation secondary quantity
<dt><code><em>dir</em></code>
<dd>is the number of directory blocks
<dt><code><em>dsorg</em></code>
<dd>is the dataset organization: <code>PS</code>, <code>PO</code>,
<code>DA</code>, or <code>IS</code>,
<dt><code><em>recfm</em></code>
<dd>is the record format: <code>F</code>, <code>FB</code>, <code>FBS</code>,
<code>V</code>, <code>VB</code>, <code>VBS</code>,
or <code>U</code>.
<dt><code><em>lrecl</em></code>
<dd>is the logical record length
<dt><code><em>blksize</em></code>
<dd>is the block size
<dt><code><em>keylen</em></code>
<dd>is the key length
</dl>
<p>All parameters except dsname and method are optional.
Defaults of zero are supplied for DCB parameters. For datasets
loaded with the XMIT method, the DCB parameters are taken from
the unloaded PDS, and the minimum space allocation required to
load the dataset is used unless a larger quantity is specified.
If space allocation is omitted, the default is TRK 1 0 0.
If CYL is specified without any primary quantity then the default
space allocation is 1 cylinder or the minimum number of cylinders
required to load the dataset, whichever is larger.
<h4>Examples</h4>
<p>
<b>[1]</b> To create a 2314 volume in a file called <b><i>sysres.230</i></b>
using the control file <b><i>sysres.plf</i></b> with message level 2:
<pre><code>dasdload sysres.plf sysres.230 2</pre></code>
<p>
An example control file is shown below:
<code><pre>#
# Pack layout file for MFT system residence volume
#
sysres 2314 * ieaipl00.rdr
sys1.parmlib xmit /cdrom/os360/reslibs/parmlib.xmi
sys1.imagelib xmit /cdrom/os360/reslibs/imagelib.xmi
sysctlg cvol trk 1 0 0 ps f 256 256 8
sysvtoc vtoc trk 5
sys1.logrec dip trk 1 0 0
sys1.nucleus xmit /cdrom/os360/reslibs/nucleus.xmi cyl
sys1.svclib xmit /cdrom/os360/reslibs/svclib.xmi cyl
sys1.sysjobqe empty cyl 2 0 0 da f 176 176 0
sys1.dump empty cyl 10 0 0 ps u 0 3625 0
</pre></code>
<p>
<b>[2]</b> To create a compressed 3390-3 volume in a file called
<b><i>linux.500</i></b> containing a bootable linux system for
linux/390 installation using the control file <b><i>linux.prm</i></b>:
<pre><code>dasdload -z linux.prm linux.500</pre></code>
<p>
An example control file is shown below:
<code><pre>
#
# Build a bootable linux disk
# [Note: the dataset names (sys1.linux. ...) are hard-coded in
# linuxipl.obj and cannot be changed without rebuilding it]
#
linux 3390-3 * linuxipl.obj
sys1.linux.parmfile SEQ images/redhat.prm trk 1 0 0 ps fb 1024 1024
sys1.linux.tapeipl.ikr SEQ images/kernel.img trk 200 0 0 ps fb 1024 1024
sys1.linux.initrd SEQ images/initrd.img trk 200 0 0 ps fb 1024 1024
</pre></code>
<h3>Fixing the XCTL tables in SVCLIB</h3>
<p>
On an OS/360 system, the Open/Close/EOV modules in SYS1.SVCLIB have
XCTL tables embedded within them. These tables contain TTRs
pointing to other modules, and these TTRs need to be adjusted after
loading SVCLIB to DASD. OS/360 provides a program called IEHIOSUP
to perform this function, but the catch is that you can't run
IEHIOSUP until you have the system up and running, and you can't IPL
until you have fixed the XCTL tables.
To solve this problem, Hercules provides a program called dasdisup
which can be run from the Unix command line after running dasdload.
<p>
The format of the dasdisup command is:
<p><code>dasdisup <em>outfile [sf=shadow-file-name]</em></code>
<p>
where
<dl>
<dt><code><em>outfile</em></code>
<dd>is the name of the DASD image file to be updated
<dt><code><em>shadow-file-name</em></code>
<dd>(optional) is the name of the associated shadow file
as specified in the Hercules config file
</dl>
<p>
<em>Note: do not use this procedure except on OS/360 IPL volumes;
other operating systems do not have XCTL tables.</em>
<h2>Other DASD utilities</h2>
<p>
These programs can be used to extract data from CKD DASD images
by means of commands issued at the Unix shell prompt.
<h3>DASDLS - List datasets on volume</h3>
<p>
DASDLS, written by Malcolm Beattie, is a command to let you list the
names of the datasets contained in disk images.
<p>
The command format is:
<br><code>
dasdls <em>ckdfile [sf=shadow-file-name] ...</em>
</code></br>
where <em>ckdfile</em> is the name of a Unix file containing a CKD volume
and <em>shadow-file-name</em> (optional) is the name of the associated
shadow file.
<h3>DASDCAT - Display PDS members</h3>
<p>
DASDCAT, written by Malcolm Beattie, is a command to let you read
datasets from disk images.
<p>
The command format is:
<br><code>
dasdcat -i <em>ckdfile [sf=shadow-file-name] dsname1 dsname2 ...</em> -i <em>ckdfile2 dsname10 ...</em>
</code><br>
where <em>ckdfile</em> is the name of a Unix file containing a CKD volume,
<em>shadow-file-name</em> (optional) is the name of the associated shadow file,
and <em>dsname</em> can be a plain (non-partitioned) dataset name
(which is currently not handled) or of the form <em>pdsname</em>/<em>memname</em>
where <em>memname</em> can be:
<ul>
<li><b>PDS member name</b> (automatically uppercased),
optionally followed by ":" and flags "a" or "c".
<ul>
<li>"c" means (c)ard images and turns a PDS members with a block
size that's a multiple of 80 into multiple newline separated
lines of 72 characters with EBCDIC converted to ASCII and with
sequence numbers chopped off.
<li>"a" means (a)sciify the member (but don't chop off sequence
numbers or do the card image thing).
</ul>
<li><b>?</b> (don't forget to quote it to avoid the shell globbing it)
to list the names of all PDS members instead of outputting their
contents.
<li><b>*</b> (again, quote it or backwhack it to avoid it being a glob)
to output all members of the PDS instead of just a named one.
This can optionally be followed with colon-then-flags, as above.
Each member is preceded with a line "> Member: memname" and, if
the "c" for card-images flags is used, each line of the members'
contents is preceded with "| " to guarantee it can be
distinguished from contents.
</ul>
<h5>Examples</h5>
<pre><code>
% dasdcat -i mvtres.350 sf= mvtres_1.350 'sys1.parmlib/?'
ieabld00
ieaige00
ieaigg00
ieaigg01
iearsv00
ikjprm00
lnklst00
presres
smfdeflt
% dasdcat -i mvtres.350 sys1.parmlib/smfdeflt:c
OPT=2, SYSTEM,JOB AND STEP DATA COLLECTION
EXT=YES, USER EXITS ARE TO BE TAKEN
JWT=15, MAXIMUM CONTINUOUS WAIT TIME IS 15 MINS.PER STEP
BUF=400, A MINIMUM 400 BYTE BUFFER IS DEFINED
SID=6A, SYSTEM ID IS 6A
MDL=65, MODEL IS MOD 65
OPI=YES, PERMIT OPERATOR INTERVENTION
MAN=ALL, RECORD USER AND SYSTEM RECORDS
PRM=(,282,NL) SYS1.MAN ALLOCATED TO NON-LABELED TAPE
% dasdcat -i mvtres.350 sys1.help/\*:c
> Member ACCOUNT
| )S SUBCOMMANDS -
| ADD/A,CHANGE/C,DELETE/D,LIST/L,LISTIDS/LISTI,HELP/H,END
| )F FUNCTION -
| THE ACCOUNT COMMAND PROCESSOR INVOKES THE CONVERSATIONAL PROGRAMS
...
> Member ALLOC
| )F FUNCTION -
| THE ALLOCATE COMMAND DYNAMICALLY DEFINES AND ALLOCATES A DATA SET
| WITH OR WITHOUT AN ATTRIBUTE LIST OF DCB PARAMETERS
| )X SYNTAX -
| ALLOCATE DATASET('DSNAME'/*) FILE('DDNAME')
...
</code></pre>
<h3>DASDPDSU - Unload PDS members</h3>
<p>
DASDPDSU is a command which unloads PDS members from a disk image
and copies each member to a file <em>memname</em>.mac in the current
working directory.
<p>
The command format is:
<br><code>
dasdpdsu <em>ckdfile [sf=shadow-file-name] pdsname</em> [ascii]
</code></br>
where <em>ckdfile</em> is the name of a Unix file containing a CKD
volume, <em>shadow-file-name</em> (optional) is the name of the associated
shadow file, and <em>pdsname</em> is the name of a PDS on that volume.
If the optional <b>ascii</b> keyword is specified, the members will be
unloaded as ASCII variable length text files. Otherwise the members
are unloaded as fixed length EBCDIC binary files.
<p><center><hr width=15% noshade></center>
<p>
If you have a question about Hercules, see the
<a href="hercfaq.html">Hercules Frequently-Asked Questions</a> page.
<p><center><hr width=15% noshade>
<a href="hercinst.html"><img src="gifs/back.gif" border=0 alt="back"></a>
</center>
<small>
<p>Last updated 30 May 2003
</BODY>
</HTML>
Reference in New Issue View Git Blame Copy Permalink