Creating DASD¶
Using pre-built DASD images¶
IBM distributes pre-built OS/390 and z/OS systems on two different CD-ROM packages:
The OS/390 and z/OS Application Development CD (ADCD)
available only to members of IBM PartnerWorld for Developers, and
The OS/390 and z/OS DemoPkg
available only to IBM employees and qualified IBM Business Partners.
Both of these packages contain pre-built DASD image files which simply need to be unzipped onto your hard drive. The unzipped images can be directly read by Hercules. Be aware, however, that you cannot use the ADCD images because the PartnerWorld scheme requires you to purchase or lease an IBM approved machine in order to obtain the ADCD, and the software on the ADCD is licensed for use only on the machine that it was shipped with. See IBM’s “Enterprise server solutions” web page for more information.
If you want Hercules to be an approved machine so that you can use the ADCD, then I suggest you lobby IBM Developer Relations at the address given on their web page. Different rules apply to the OS/390 and z/OS DemoPkg CD which is available only to IBM employees and business partners. If you fall into this category then you probably know what the rules are. I don’t. :-(
Creating, formatting, and loading DASD volumes¶
Creating an empty DASD volume¶
The dasdinit or dasdinit64 utility must first be run from the Unix shell prompt to create a file containing an empty DASD volume. If you are going to be creating compressed dasd images, it is recommended that you use the new dasdinit64 utility as it is able to create CCKD64 format dasd images which are highly advantageous over the old 32-bit CCKD images that plain dasdinit creates.
Usage: dasdinit64 [-options] filename devtype[-model] [volser] [size]
Builds an empty dasd image file
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 a large (uncompressed) dasd file (if supported)
-a build dasd image file that includes alternate cylinders
(option ignored if size is manually specified)
-r build 'raw' dasd image file
(no VOL1 or IPL track)
-b make wait PSW in IPL1 record a BC-mode PSW
(default is EC-mode PSW)
-m enable wait PSW in IPL1 record for machine checks
(default is disabled for machine checks)
-linux null track images will look like Linux dasdfmt'ed images
(3390 device type only)
filename name of dasd image file to be created
devtype CKD: 2305, 2311, 2314, 3330, 3340, 3350, 3375, 3380, 3390, 9345
FBA: 0671, 3310, 3370, 9313, 9332, 9335, 9336
model device model (implies size) (opt)
volser volume serial number (1-6 characters)
(specified only if '-r' option not used)
size number of CKD cylinders or 512-byte FBA sectors
(required if model not specified else optional)
The current list of device types and models supported is:
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
3390-27 32760 3
3390-54 65520 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
9313 [*]
9313-1 246240
9332 [*]
9332-200 360036
9332-400 360036
9332-600 554800
9335 [*]
9335-1 804714
9336 [*]
9336-10 920115
9336-20 1672881
9336-25 1672881
0671-08 513072
0671 574560
0671-04 624456
[*] size may be specified else size defaults to the first listed model.
Volumes exceeding 2GB¶
For regular (uncompressed) CKD volumes which exceed 2GB in size (such as the 3390-3 and larger models) – and for which the -lfs parameter was not specified – the DASDINIT / DASDINIT64 program creates multiple files by appending the characters _1, _2, _3 etc. to the file name specified on the command line. These characters are inserted before the first dot (.) after the last slash (/). 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. “filename_1”).
If the -lfs option is specified however, then the output file is a single large file which can be as large as your system supports.
The DASDINIT / DASDINIT64 program cannot create FBA volumes exceeding 2GB unless the -lfs parameter is specified and large file size is supported on your platform..
Examples¶
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 work01.151:
dasdinit64 -a work01.151 3330-1 work01
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 dosvs34.24f:
dasdinit64 -a -bz2 dosvs34.24f 3350-1 sysres
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 mini.work02.140:
dasdinit64 mini.work02.140 3370 work02 100000
To create a 3390 model 3 (triple density) CKD volume of 3339 cylinders with volume serial number WORK03:
dasdinit64 triple.a88 3390-3 work03
Because this volume exceeds 2GB, DASDINIT / DASDINIT64 will create two files with triple_1.a88 containing cylinders 0-2518 and triple_2.a88 containing cylinders 2519-3339. However, if you instead specify the ``-lfs`` option:
dasdinit64 -lfs triple.a88 3390-3 work03
then DASDINIT /DASDINIT64 will create a single file triple.a88 containing all the cylinders. Your platform must support large file sizes to specify the -lfs option.
Formatting the empty DASD volume¶
After creating a DASD volume you can format it with a program such as standalone IBCDASDI or ICKDSF.
Here is an example of the IBCDASDI control statements required to initialize a 3330 volume:
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
To run IBCDASDI, place the above statements in a file called init3330.txt and start Hercules in S/370 mode with a configuration file containing these statements:
CPUSERIAL 001234
CPUMODEL 3145
MAINSIZE 2
CNSLPORT 1052
ARCHLVL S/370
0009 1052
000A 1442 ibcdasdi.rdr
000C 1442 init3330.txt
0151 3330 work01.151
After IPLing from card reader device 00A, connect a telnet client to port 1052, and press enter. At the IBCDASDI prompt, enter the command:
input=1442 00c
Loading the new DASD volume¶
Next, you need to create a full volume dump file of your chosen mainframe dasd volume (converting it to AWSTAPE format) by using the tapeconv.jcl job in the Hercules source directory. You would run this JCL on your mainframe.
The resulting AWSTAPE mainframe file can then be downloaded in binary format to your PC where Hercules is running, where it can then be defined as a Hercules virtual tape drive in your Hercules configuration file.
A standalone program could then be IPLed on Hercules to restore the volume image from the virtual tape onto the formatted Hercules DASD volume.
Building a DASD volume from unloaded PDS files¶
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.
The format of the dasdload / dasdload64 command is:
dasdload [options] ctlfile outfile [msglevel [maxdblk maxttr maxdscb]]
dasdload64 [options] ctlfile outfile [msglevel [maxdblk maxttr maxdscb]]|
where [options] can be:
- -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.
- -a
Build dasd image file that includes alternate cylinders.
- -b
For a volume without IPL text, make the wait PSW written to the IPL1 record a BC-mode PSW. The default is to make the wait PSW an EC-mode PSW.
- -m
For a volume without IPL text, make the wait PSW written to the IPL1 record enabled for machine checks. The default is to make the wait PSW disabled for machine checks.
ctlfile is the name of the control file which specifies the datasets that are to be loaded onto the newly-created volume
outfile is the name of the DASD image file to be created
msglevel is an optional number from 0 to 5 (default is 1) which controls the level of detail of the messages issued during the load.
maxdblk is the optional maximum number of DBLK table entries or 0 to use the default
maxttr is the optional maximum number of TTR table entries or 0 to use the default
maxdscb is the optional maximum number of DSCB table entries or 0 to use the default
--option1: value1--option2: value2--option3: value3
Note that dasdload’s default for the wait PSW written to the IPL1 record have changed from earlier releases of Hercules. In the past, the wait PSW created by dasdload on volumes without IPL text was a BC-mode PSW enabled for machine check interrupts. The current default for the wait PSW is EC-mode, disabled for machine checks. To obtain the earlier behavior, run dasdload with the “-b” and “-m” flags.
Control file¶
The control file required by the dasdload /dasdload64 program is an ASCII text file consisting of a volume statement followed by one dataset statement for each dataset to be created.
The format of the volume statement is:
``volser devtype``[-``model``] [``cyls``[``ipltext``] ]
where:
- ``volser``
is the volume serial number for the newly-created volume
- ``devtype``
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 / dasdload64 program. Model may be specified like dasdinit / dasdinit64 above.
- ``cyls``
is the size of the new volume in cylinders. If ``cyls`` is coded as
*or as0or is omitted, then the default size for the device type and model is used.- ``ipltext``
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.
The format of a dataset statement is:
``dsname method units pri sec dir dsorg recfm lrecl blksize keylen``
where:
- ``dsname``
is the dataset name
- ``method``
is the dataset loading method which can be one of the following:
XMIT``filename``the dataset is loaded from an unloaded PDS created by the TSO XMIT command
SEQ``filename``the sequential dataset is loaded from a binary file. ascii/ebcdic translation is not currently supported. Also, the dsorg must either be PS or DA and recfm must either be F or FB.
EMPTYthe dataset is initialized with an end of file record (if DSORG is PS) or an empty PDS directory (if DSORG is PO)
DIPthe dataset is initialized with a LOGREC header record
CVOLthe dataset is initialized as an OS SYSCTLG containing the minimum entries needed to IPL an OS/360 system
VTOCspecifies 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.
- ``units``
is the space allocation units:
TRKorCYL.- ``pri``
is the space allocation primary quantity
- ``sec``
is the space allocation secondary quantity
- ``dir``
is the number of directory blocks
- ``dsorg``
is the dataset organization:
PS,PO,DA, orIS,- ``recfm``
is the record format:
F,FB,FBS,V,VB,VBS, orU.- ``lrecl``
is the logical record length
- ``blksize``
is the block size
- ``keylen``
is the key length
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.
Example 1:¶
To create a 2314 volume in a file called sysres.230 using the control file sysres.plf with message level 2:
dasdload64 sysres.plf sysres.230 2
An example control file is shown below:
#
# 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
Example 2:¶
To create a compressed 3390-3 volume in a file called linux.500 containing a bootable linux system for linux/390 installation using the control file linux.prm:
dasdload64 -z linux.prm linux.500
An example control file is shown below:
#
# 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
Fixing the XCTL tables in SVCLIB¶
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-22 situation 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 dilemma, Hercules provides a program called dasdisup which can be run from the Unix command line after running dasdload / dasdload64.
The format of the dasdisup command is:
dasdisup``outfile [sf=shadow-file-name]``
where
- ``outfile``
is the name of the DASD image file to be updated
- ``shadow-file-name``
(optional) is the name of the associated shadow file as specified in the Hercules config file
Note: do not use this procedure except on OS/360 IPL volumes! Other operating systems do not have XCTL tables!
Other DASD utilities¶
These programs can be used to extract data from CKD DASD images by means of commands issued at the Unix shell prompt.
DASDLS, written by Malcolm Beattie and enhanced by others, is a command to let you list the names of the datasets contained in disk images.
The command format is:
dasdls [-option [-option ... ]] ckdfile [sf=sfile] [...]
where ckdfile is the name of a Unix file containing a CKD volume and sfile (optional) is the name of the associated shadow file.
-option can be:
- ``-hdr``
show column headers
- ``-dsnl [=n]``
restrict dsname width
- ``-info``
show F1 info
- ``-caldt``
calendar date format
- ``-refdt``
show last-reference date
- ``-expdt``
show expiry date
- ``-yroffs [=n]``
year offset
Note: Multiple images can be processed in the same run, but options must be specified ahead of each image.
DASDCAT, written by Malcolm Beattie, is a command to let you read datasets from disk images.
The command format is:
dasdcat -i``ckdfile [sf=shadow-file-name] dsname1 dsname2 …``-i``ckdfile2 dsname10 …``
where ckdfile is the name of a Unix file containing a CKD volume, shadow-file-name (optional) is the name of the associated shadow file, and dsname can be a plain (non-partitioned) dataset name (which is currently not handled) or of the form pdsname/memname where memname can be:
PDS member name (automatically uppercased), optionally followed by “:” and flags “a” or “c”.
“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.
“a” means (a)sciify the member (but don’t chop off sequence numbers or do the card image thing).
? (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.
* (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.
Examples:¶
% 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')
...
DASDPDSU: Unload PDS members¶
DASDPDSU is a command which unloads PDS members from a disk image and copies each member to a file memname.mac in the current working directory.
The command format is:
dasdpdsu``ckdfile [sf=shadow-file-name] pdsname``[ASCII] [odir]
where ckdfile is the name of a file containing a CKD volume, shadow-file-name (optional) is the name of the associated shadow file, and pdsname is the name of a PDS on that volume. If the optional ASCII 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. The optional odir parameter is the name of the directory where the output files should be placed. Otherwise if not specified they are created in the current directory.
If you have a question about Hercules, see the Hercules Frequently-Asked Questions page.