Please navigate by using our sitemap if a new menu does not appear shortly.

What has changed in revisions from 1.0 to 4.13       (click here to page down directly to revision 4.13)

Revision 1.0

Beta Code:

RamDisc Classic code revision 1.9 transferred to RamDisc Plus

Revision 1.1

Beta Code:

Added restricted maximum memory configuration license code.

Revision 1.2

Beta Code:

Added option to save the data to the alternate disk if two hard disk
drives are configured and the volume group is deleted or modified.

Revision 1.3

Beta Code:

This applies to the RamDisc Classic portion only...

The ramdisc filesystem can be deallocated to free up memory to be used
by other applications without a system reboot. The filesystem capacity
may be resized with a system reboot. In both of the above cases ALL OF
THE DATA in the previous ramdisc filesystem will be lost.  This option
is disabled if the following condition occurs:

The ramdisc mount pathname was changed without a system reboot.

Revision 1.4  (see revision 1.7)

Beta Code:

Added new RamDisc Environmental variable RAM_MAXIMUM...(See section 10)
If not specified as an environmental variable then the following default
values will be used:

10.00 - 10.20 : 768 MB
11.00 32bit   : 768 MB
11.00 64bit   : 62404 MB aka 60GB
11.11 32bit   : 768 MB
11.11 64bit   : 62404 MB aka 60GB

Any time the ramdisc capacity is increased without a reboot a warning will
be displayed indicating the possibility of the following system panic:

panic: kalloc out of kernel virtual space

The most likely times of occurrence:

1. The ramdisc is near the upper limit for maximum capacity (32 bit)
2. The system has other programs/applications previously executed.

* This is not a problem on 64bit mode operating systems such as 11.00

Another change was to reboot the system twice after an HP-UX 11 kernel
rebuild to properly install the kernel modules. This is the default.

Revision 1.5

Beta Code:

Memory license is now displayed in the upper right hand corner of the
main menu after the license key is correctly entered.

Added an End User License Agreement notice when starting up for the
first time or when updating to a new revision of RamDisc.

Revision 1.6

Public Release:

When increasing logical volumes (providing there is enough free space)
the filesystem is extended by using /sbin/extendfs instead of deleting
and re-creating the volume. No loss of data for volume increase.

Revision 1.7 (see revision 2.0)

Public Release:

The copy data menu for the RamDisc Classic portion of RamDisc Plus would
pop up twice after data has been entered. This is now corrected.

Reduced the default RAM_MAXIMUM variable from 768mb to 704mb for systems
that are not running a 64bit OS. This was done to reduce the chance of a
system panic: kalloc out of kernel virtual space.

This is NOT an issue when running a 64bit OS

The new RAM_MAXIMUM variable defaults:

10.00 - 10.20 : 704 MB
11.00 32bit   : 704 MB
11.00 64bit   : 62404 MB aka 60GB
11.11 32bit   : 704 MB
11.11 64bit   : 62404 MB aka 60GB

Revision 1.8

Beta Code:

Added RamDisc Ram Only package.  This package is like RamDisc Classic in
function with the additional option of specifying multiple mount points
using LVM (Logical Volume Manager). Reads and Writes are as fast as main
memory and backups may be performed using the RamDisc backup utility.
Note: The memory used in Ram Only volumes can only be deallocated after
LVM volume group removal and then a system reboot.

Revision 1.9 (see revision 2.0)

Beta Code:

All of the RamDisc Classic functions including backup utilities are now
rolled into RamDisc Plus and the maximum memory for Classic is now 4096.

Added "Include in the backup list" feature to backup only the selected
directories in the Copy Data menu.

The lowest capacity ramdisc mirrored to a hardrive is now 48 MB.  There
may be "Data page fault" panics while trying to rebuild  the mirror set
if the volume was sized at 32MB. Making the minimum value to 48MB makes
this panic go away. If you don't have 48MB available but you do have at
least 8MB then you will be asked if you want  to make a Ram Only volume
instead. This volume does not cause the system to panic because it gets
created upon every system boot instead of restored from disk drives.

Revision 2.0

Public Release:

Release 2 revision 0

Added different license code as compared to release 1

After using ramdisc on a V2250 with 12GB of usable ram is was discovered
that the maximum single ram device file can only be 2044MB instead of the
calculated figure of 4092MB so now the RAM_MAXIMUM variable is now set to
30660MB for 64bit. Do not go above this value if you are manually setting
this variable.

The new RAM_MAXIMUM variable defaults:

10.00 - 10.20 : 704 MB
11.00 32bit   : 704 MB
11.00 64bit   : 30660 MB aka 30GB
11.11 32bit   : 704 MB
11.11 64bit   : 30660 MB aka 30GB

Revision 2.1

Included a non-graphical (text only version) menu utility. This is the
default menu system when no X11 connection exists.    Useful in server
configurations or telnet sessions. Supported in all common HP terminal
types including VT100 telnet sessions from a PC to the HP system.

Revision 2.2

Updated the way /sbin/ramdisc finds and checks the ramdisc driver based
upon several HP-UX software revisions. (Required for HP-UX 11i [11.11])

Revision 2.3

Several people mentioned that they  were going  to use the RamDisc as a
RAW character device (no format or mount).  You now have a third choice
when creating logical volumes to use the volume as a  RAW  device. Data
may be copied using the dd command  (with dd options)  upon system boot
and may be backed up with the  RamDisc  backup utility.  Another change
is to  backup all data marked with "Add to the backup list" button when
executing the system shutdown command (goes through shutdown checklist)

Please note:  The RAW character device option is a beta release. Any
              future modifications and support for  this option will
              be continued in release 3

Files modified to support the RAW character device:

Revision 2.4

LVM patches for HP-UX 11.11 (11i) and  possibly 11.00 requires that the
device files for any physical volume(s) in the Volume Group must reside
in the /dev/dsk and /dev/rdsk directories.  Earlier revision of RamDisc
placed the ramdisc device files in  /dev/ram and  /dev/rram directories
using the c0txd0 naming convention. Now the ramdisc device files are in
the  /dev/dsk and  /dev/rdsk directories using a new naming convention.
The ramdisc device files use "crtxd0" so that the pathnames of physical
disk drives and the ramdisc drive do not cross.

vxfs (JFS) filesystems require that the block and raw device be present
during mount.  Earlier versions of HP-UX  only needed the block device.
The Classic portion of Plus was modified to provide both  raw and block
devices during every Classic operation of the ramcom utility.

Revision 2.5

Added a demo license coding with a maximum of 2000MB (2GB).  Originally
only RamDisc Classic could be used as a demo.  If the demo expires upon
reboot, the RAM portion of the volume group will be offline.  To reduce
downtime, there is an  automated method  of extending your demo license
by using  and enter your LAN id
and your email address and a extended demo license will be sent to your
email containing an new expiration date and a new code.  The expiration
date is a negative Julian date to indicate that this is a demo license.

Revision 2.6

Completed the non-graphical text dialog menus and major notes/warnings.

Revision 2.7

/sbin/ramdisc:  Modified to determine  processor type and report to the
user to upgrade to release 3 if the processor is Itanium. A better scan
of disk devices to weed out removable media devices.  A safety check to
not report VxVM disks  (Veritas Volume Manager)  as useable for RamDisc
during the disk scan.  If the vginfo command is out of date then do not
continue to scan for useable disks since it doesn't pick up VxVM types.
/etc/ramvgrc:  Modified to mount the /stand and /usr directories if the
OS is hp-ux 11.20 or greater.  The /sbin/ramcom utility may need to use
shared libraries upon boot so the  /usr  logical volume must be online.

Revision 2.8

Modified demo license to  allow maximum memory usage based upon kernel.

Revision 2.9

This update is the same update from release 3 (see revision 3.2 below).

Revision 3.0

Release 3 revision 0

Added different license code as compared to release 1 and 2.

Included support for the IA-64 processor from Intel/HP.  This processor
uses a different OS than the HP PA-RISC processors. Made changes to the
/sbin/ramdisc program to determine processor type so the correct driver
is added to the kernel. The RAW character device option mentioned above
in rev 2.3 is really an enhancement for this revision.  This option was
included in rev 2.3 as a beta release.  Any future modifications to the
RAW character device option will be made in revisions greater than 3.0

Revision 3.1

In the past revisions, only the memory used in the  Classic  mode could
be deallocated when no longer required without rebooting.   When an LVM
group is no longer needed,  the memory that was used is now deallocated
when the volume group is removed with the /sbin/ramdisc utility without
having to reboot the system. Any  new  or  modified  volume groups will
still have to be rebooted insure that there is enough memory available.

Revision 3.2

Added the detection of  system VGs  to avoid conflicting minor numbers.
In the past,  drives larger than 73GB could not  be used for  mirroring
because of the default PE size of 4MB was not large enough.   Any disks
used for mirroring will now be calculated for larger PEs and the max PE
will allow replacement drives 15% larger than the highest capacity disk
currently in the ramdisc VG. The VG and LV capacity entries are checked
with the volume group PE size for proper LVM configuration. One PE will
be deducted from the  RamDisc Volume Group  capacity when mirroring for
proper rebuilding of the ramdisc drive. (LVM vgsync function)

Revision 3.3

Added support for 11.22 software. The vginfo command was different than
the 11.20 version (SAM uses this command to find LVM information.)

Revision 3.4

The vginfo command used by SAM will display the RamDisc volume group if
other disk volume groups are created afterwards in  HP-UX version 10.20

RamDisc Plus logical volume filesystems have the largefiles option  now
enabled by default on lvols greater than 2000MB (2GB).  To disable this
auto feature, create the filename /etc/ram-nolargefiles ...  If an lvol
was created less than 2GB  and then extended in the ramdisc menus,  you
must use the  fsadm  command if you want largefiles (over 2GB) support.

Revision 3.5

Added support for 11.23 software (11iv2)  using the kcmodule command to
verify the static ramdisc module is in the kernel.   No kernel rebuilds
are required for 11.23,  however,  a reboot is still required to create
new ram volume groups. The lockable memory formula was revised to allow
a larger amount lockable memory for the system to boot the 11.23 OS.


Disks removed from RAM VG mirrors are now formatted with vxfs (JFS)
instead of HFS unless HFS is listed in /etc/default/fs

Lockable memory amount is obtained from  /var/adm/syslog/syslog.log
if dmesg is full of errors such as "/tmp: vxfs file system full".


Change rights of vginfo.11.22 file in the /usr/local/ramdisc directory
and changed vginfo error output to /dev/null. Starting with 11.11, the
vginfo -l option requires a -t LVM option as well - Fixed when needed.

If a ramdisc device file was in the /dev/rdsk directory on system boot
for 11.20,  the lockable memory was reduced by the  capacity of device
file.  This does not occur for 11.23  (and perhaps patched versions of
11.20).  The solution in previous revisions  was to add the capacities
to the lockable memory amount. This calculation method has been turned
off by default.  If the  lockable memory is reported incorrectly  upon
system reboot, simply create the filename /etc/ramlock to enable again

Revision 3.6

In revision 3.6 and later the Plus portion (LVM configuration) may use
the Classic device file to create  LVM volume groups  and increase the
max capacity to 32704 MB.  This only applies to  large memory  systems
exceeding 32GB.  Using this option for the large LVM creation disables
the use of the RamDisc Classic  menus until  reconfiguration.  Earlier
revisions could use 32704 MB but  only  when combined with the maximum
of 30660 MB with RamDisc Plus and 2044 MB with RamDisc Classic.

The find command no longer supports the -mountstop option. Modified to
use the  -xdev  option instead in the RamDisc copy and backup scripts.

Revision 3.7

When compiling the new kernel,  the krm module  needs to be recompiled
as well on HP-UX 11.x revisions.  Added to compile when needed on 11.x
When scanning system volume group files  (/dev/vgxx/group) , any minor
number that contains an alpha character outputs "a: bad number" in the
ramdisc utility while gathering volume group information.  The ramdisc
utility will now ignore any alpha characters within the group file.

Revision 3.8

Changed support for 11iv2 (11.23) software (see revision 3.5). PA-RISC
systems need a revised driver added to kernel, whereas Itanium systems
require the ramdisc driver remain static before and after installation
or removal. Some menu enhancements were also included.


If the  -install option was entrered after the initial installation of
RamDisc Plus with 11.23 Itanium, the install utility would replace the
driver already in the kernel.  New code does not allow this to happen.

Revision 3.09

Added a third chance to determine the lockable memory amount. If dmesg
fails, then it is read from /var/adm/syslog/syslog.log ... If for some
reason, syslog does not contain the memory information, then read from
the /etc/ramvgrc.loc file that was created upon system boot.

Revision 3.10

Potential customers asked for a method to operate  RamDisc Plus with a
global site license. This would be used in critical operations whereas
a replaced FRU (component)  would not cause the license to be invalid.
This license coding is installed in any future release of RamDisc Plus

Revision 3.11

Code was modified to support RamClustered volume groups (beta release)
Optional software MirrorDisk/UX and MC/ServiceGuard  must be installed
and configured in order to support the failover of a RamCluster. There
are step by step instructions in the  /usr/local/etc/ramdisc directory
(README.ramcluster) to create the RamCluster and include scripts to be
added to the customer defined run and halt package functions.

Revision 3.12

Modified  single disk  mirror warnings to be ignored when a LUN is the
device that will be mirrored to the ramdisc. Now the RamCluster volume
group ramdisc device pathnames do not have to match  for both systems.
During initial configuration,  RamDisc searches for  available ramdisc
devices to create the Ram Volume Group.  The list of devices used  may
be different pathnames between the two systems in the cluster.  In rev
3.11 the complete list of pathnames must be the same so the RamCluster
volume group could be imported.  The vgcfgrestore command allows for a
restore from an old PV path to a current PV path  so changes were made
to support flexible ramdisc pathnames.   Although the pathnames may be
different, the memory amount used in each RamCluster Volume Group must
remain the same. The README.ramcluster file has new instructions.

To avoid conflicting minor numbers with system VGs, changes were added
back in rev 3.02 to avoid using the same minor numbers in the VG group
file. However, the vgdisplay -v command was used to obtain the list of
volume groups configured.  Any volume group not activated would not be
included in the list allowing a potential duplicate minor number. This
problem would likely occur in a ServiceGuard environment. Now the list
of configured volume groups is obtained from the /etc/lvmtab file.

Revision 3.13

Updated license and technical request  contact information.  There are
no other changes in 3.13,  however, revsion 3.xx users should download
this revision since the website information included in  revision 3.xx
is now considered obsolete.

Revision 4.00

There are three major changes for release 4 and a few minor changes:

    RamDisc capacities greater than 32GB are available for HP-UX 11.23
    (Itanium or PA-RISC).   Capacities beyond 512GB may be configured.

    RamDisc Classic using 11.23 (11i v2)  may be configured up to 64GB
    using a single device file (non-LVM).

    RamClustered Volume Groups using MC/ServiceGuard and MirrorDisk/UX
    are now supported in this release. (see revision 3.11 above)

    The mirror option  now deallocates enough hard disk space to cover
    all current (and future) lrams.

    The raw and block device file used for RamDisc Classic now remains
    in the /dev/rdsk and /dev/dsk directories.

    Include the strings of  /etc/lvmtab  during drive scans to include
    the physical volumes that are currently offline.

    The license coding has changed so customers with a current support
    contract may upgrade RamDisc Plus and convert to release 4 license
    by going to    

Revision 4.01

Updated license and technical request  contact information.  There are
no other changes in 4.01,  however, revsion 4.00 users should download
this revision since the website information included in  revision 4.00
is now considered obsolete.

Revision 4.02

In previous revisions, if system volume groups used group control file
1 - 9 then no  RamDisc volume groups could be created until one of the
system volume groups was moved to a higher value.  A new option called
Expanded Ram Volume Group was created to allow the ramdisc menus to be
able to configure a RAM volume group (group control file) greater than
number 9 and warn to adjust the tuneable parameter "maxvgs" before the
next system reboot (if needed).  There are  two methods to enable this
new option.  The first method is exceeding a threshold variable within
the /sbin/ramdisc program currently set to if only two RAM Volumes are
available for configuration  due to other volume groups on the system.
The other method is a command line option. By entering -expramvg as an
additional command line argument,  you can bypass the built-in option.

Added a selection of mount options for HFS and VxFS.  In the directory
/usr/local/etc/ramdisc/fs,  there are two files called fsopts.vxfs and
fsopts.hfs that contain variable type mount options.  You may add your
own options as well. New versions of these files will be stored in the
/usr/local/etc/ramdisc/fs/newconfig  directory so your selections will
not be overwritten when upgrading.  These files also include the large
files mount option  that has been a problem in  previous revisions for
VxFS  filesystems when the defaults  parameter was used in /etc/fstab.

Revision 4.03

In previous revisions, a system being recovered from  Ignite/UX  would
fail when trying to import any RamDisc volume group. Changes were made
to /etc/ramvgrc to detect the initial boot from the Ignite/UX  restore
and allocate the ram required then let Ignite/UX run the vgimport.

There have been a few customers that had to perform the vgscan command
due to a problem with non-RamDisc volume groups. This command will try
to rebuild the LVM structure for the system by scanning physical disks
attached. However, vgscan does not check for RamDisc devices and can't
determine the correct entries for  /etc/lvmtab.  RamOnly volume groups
are corrected by rebooting the system after the vgscan command. Volume
groups that are configured as  RamDisc  mirrors require manual edit of
/etc/rc.config.d/ramvol then reboot.  Please contact Technical Support
for mirrored or RamCluster configurations for further explanation.

Two RamCluster programs (ramcluster_run ramcluster_halt) were modified
to use the mount command instead of bdf. Both scripts also include the
fuser command and works just like normal  ServiceGuard  if a directory
can't umount because of "Device Busy" the fuser command will function.

Revision 4.04

Improved code for PA-RISC using 11.23  during ramdisc installation and
detecting if the files  needed for capacities greater than 32GB are in
place. Corrected visual message for adb -k option that 11.23 displays.

Revision 4.05

Introduced auto scan and auto configuration  of alternate device paths
into the RamDisc volume groups when only a single RAID LUN  (or a disk
with more than one path)  is selected for mirroring to the ram device.

Revision 4.06

Newer versions of  Ignite/UX  would create a mapfile for vgimport that
contains the VGID of the RamDisc volume groups  causing a failure upon
import. Corrected this problem by allowing /etc/ramvgrc to remove this
entry from the mapfile before the Ignite/UX post scripts are executed.

Earlier searches for alternate paths only used the first 30 characters
of the additional inquiry bytes when diskinfo -v was used to determine
drive identity. Some RAID models use unique characters near the end of
the inquiry byte string to identify each RAID LUN.  The alternate path
search was modified to use  all of the characters  in the byte string.

Revision 4.07

Added initial support for 11.31 software (11iv3). A warning added when
scanning for disk devices to be used for LVM mirroring configurations.
Later versions of 11.23 could not obtain the maxvgs tuneable parameter
and 11.31 does not use the maxvgs parameter. The /sbin/ramdisc program
now uses several different methods to  determine the maxvgs parameter.

Revision 4.08

The vxfs option largefiles is enabled by default on VxFS 4.1 and later 
releases.  Previous versions defaulted to use the nolargefiles option.
Program modifications were made  to the vxfs format and mount options.

Revision 4.09

Systems configured with greater than  25GB  of memory handle locakable
memory amounts reserved  in a  different manner. The default amount of
lockable memory reserved is  20 percent for any system configured with
more than 25GB of memory.  However, in some cases, this value may need
to be lowered to allow users  full usage  of the RamDisc license.  The
HICAPMEMLOCK variable defines the lockable percentage amount reserved.

Environmental variables like HICAPMEMLOCK may be placed into a file in
the /etc directory called ramvgrc.env so that they may be sourced when
/sbin/ramdisc is started. Revision upgrades do not overwrite the file.

Revision 4.10

A new calculation was added to more accurately determine the amount of
free space available in  the  RamDisc configuration menus prior to the
actual creation of the RamDisc volume group upon the next system boot.

Revision 4.11

Earlier versions of RamDisc included initial support for 11iv3 (11.31)
by providing  temporary LVM commands supporting Itanium architectures.
In order to support LVM configurations of RamDisc on PA-RISC platforms
using the 11.31 Operating System,  September 2008 (or greater) BaseLVM
filesets must be installed.  The previous 11.31 BaseLVM fileset may be
upgraded using September 2008 installation disks without the need of a
complete OS reload. Upon successful upgrade of the BaseLVM product, an
Itanium system will not use the temporay fix  LVM commands anymore and
PA-RISC systems will allow LVM configurations within the RamDisc menus

This issue has been resolved with the next revision of RamDisc (4.12).

Revision 4.12

Removed enhanced patch check and warning displayed for any revision of
Operating System  not matching  the criteria for the enhanced patches.

Created a method to allow PA-RISC un-patched servers to continue usage
of RamDisc by using alternate LVM commands. An attempt will be made to
use the default LVM commands for pvcreate, vgcreate and vgextend. Upon
completion, the exit status will be check and if unsuccessful, another
set of LVM commands known to succede with RamDisc devices stored in an
alternate location will be used instead to complete the configuration.

The /tmp/ram_install has a new option -altlvm that will display a menu
of alternate LVM commands currently available within RamDisc and other
optional LVM commands that may be downloaded and placed into position.
Under most conditions, the alternate LVM commands that are included in
the standard download of the RamDisc tar file should be sufficient and
this option should only be used with direction from technical support.

Revision 4.13

When mirroring to a mass storage device within 11.31 (11iv3) an option
was added to allow the  use of persistent device special files (DSF's)
also known as "Agile" device files.  This type of device file uses the
mass storage stack's multi-pathing functionality and  will bind to the
World Wide Indentifier (WWID) of  the  mass storage device thus making
the device persistent. In other words, each LUN is uniquely identified
by a single device special file irrespective to the number of lunpaths
and independent of any physical path changes.  Legacy device files are
still supported within 11iv3 and  RamDisc will allow the configuration
of mirrored volume group using either the Agile or legacy device file.

Other pages within the AppMate website...
[Home]  [Software]  [Contact Us]  [References]  [Geek stuff]
AppMate Performance Software
207 Long Pond Rd
Danville, NH 03819
Tel: (603) 617-2833
Fax: (603) 617-2891
Tel: (888) 601-3069