Back to top
Question: I have a problem with the LCCMUSER
account doing unattended
installations with Windows98®SE.
When Windows 98 tries to logon to complete the
installation, only the letters "SER"
are in the user name field. If I prepend "lccmu"
to the User Name, then the login finishes and so does the installation.
How do I fix this?
Answer: This happens if the LCLI_REGUSER
field in a wizard profile, has more than 20 characters in it.
For now, we suggest you limit this field to 20
characters, and then use an automatic regedit from
within the CLIENT.BAT file to increase the length of the string in the registry.
If you do a manual interactive installation of Windows 98SE, the dialog box for this field allows you
to enter as many as 29 characters; but only the first 24
characters get displayed on the system property sheet,
even though all 29 are actually in the registry entry.
Back to top
Question: I am attempting
to configure a LCCM server running Windows NT®
4 SP6a and I am utilizing a Windows 2000 DHCP server. I
do not wish to use IBM's DHCP server; I would like to
utilize Microsoft's DHCP server. Under Windows NT®
4, I had to go and add PXEClient support (Option 60).
I thought that I also have to add that support into
Windows 2000. How do I go about adding this option?
Answer: In the configuration that you
describe, Option 60 must not be added to the Windows 2000 DHCP server.
Here is a quote from the LCCM online help:
"Select the DHCP/PXE Services you require from the following options:
- Install all DHCP Support
IBM's DHCP/PXE, BINL and TFTP services are installed.
Select this option if you do not have another DHCP server running in your environment.
- Install the IBM DHCP Proxy Service
IBM's Proxy, BINL and TFTP services are installed.
Select this option if there is another DHCP service
(e.g., Microsoft's) running in your environment, and it is not installed on your LCCM server.
- Install the BINL Service (Use with non-IBM DHCP services)
IBM's BINL and TFTP services are installed.
Select this option if there is another DHCP service
(e.g., Microsoft's) running on your LCCM server.
You must add PXE client option 60 to your DHCP service.
For more information, see Configuring PXE Option 60
for MS DHCP Server and your DHCP service's documentation."
It is fairly obvious that the ONLY possible setup is
(b), as LCCM cannot be installed on Windows 2000.
Hence, it means that the LCCM server must be separate
from the DHCP server, as the latter is on a Windows 2000 box.
A BINL service is required in order to use LCCM, and
the service MUST be running on the LCCM server.
Option 60, if configured on the standard DHCP
server tells the client that the BINL service is on the
SAME server, which, in this case, it is obviously not.
Granted that Windows 2000 has a BINL service, using
it in conjunction with LCCM is NOT a supported option, so we cannot teach you how to make it work.
The IBM DHCP Proxy service is merely a redirector,
but it must be running if you want to use LCCM in this
configuration. It requires NO setting up, as it is only
a redirector: you just need to ensure that the service is installed during LCCM setup.
Back to top
Question: There used to be a version of LCCM
that would work with
OS/2 clients, but I cannot find it anymore. I would
like to use LCCM and Drive Image to distribute OS/2
images to clients. Is there a utility that I can use
with LCCM that will allow me to access an OS/2 HPFS
drive in order to make changes to certain files?
Answer: LCCM 1.1 for OS/2 is no longer available and is no longer supported by IBM. LCCM, after
version 1.1 no longer supports OS/2 clients (HPFS). We recommend that you use CID to perform unattended
installations of OS/2 clients, or that you use a third-party application for creating OS/2 images for
sending across the network. Refer also to the entry
about using LCCM with 3rd-party imaging products.
Back to top
Question: For this client, I do
not want to deploy a clone image or do an unattended
installation. I want to keep the client as it is. How
can I have LCCM store the information on this client, so
I can update the BIOS and deploy images at a later date, but have it boot locally for now?
Answer: There are 2 alternatives.
- Create a manual profile with an empty LCI
file and no LCP file, and deploy the client to the profile.
- Assign the client to its final profile, but do NOT process it yet. Then go to the
Software page of the Client Notebook, and uncheck the
box marked Mark Client for reload, and then process the client.
Back to top
Question: I have 2 LCCM servers and a Remote
LCCM Console from which I control
the servers. Each LCCM Server has different Profiles
and Clients defined. I viewed the first LCCM Server
using the Remote Console and it looked OK. I then viewed
the second LCCM Server and it had the same screen as the
first. It appears that the Remote Console copied the
Profiles and Clients from the first to the second. When
I investigated further, the files for the Profile and
Clients still only existed on LCCM Server #1, but the
Profiles and Clients that were on LCCM Server #2 were
not referenced in any way. How can I control both
servers from the single Remote Console?
Answer: The unapproved, but nonetheless valid, way to change the server that the LCCM console
references is to force the console to ask for the server.
You do this by deleting the file LCCLIENT.INI
on the remote console. When you start the console, it
will ask for the LCCM server name. If you input the
name, LCCM will synchronize the console with the correct server.
Back to top
Question: On a client,
I have updated an IBM 10/100 EtherJet w/WOL network card
with the latest ETPFLASH (latest Intel IVY Boot Agent
v2.6 with combined PXE 2.0 (072) & RPL 1.0 code).
When I set the parameter PnP/BEV Boot to Disabled,
as per what the flash instruction says it should be set
to, I cannot boot the computer. All I get is Attaching
to Server displayed 3 times, then the system reboots
itself, and goes through the same thing endlessly.
If I change the PnP/BEV Boot to Enabled,
the system boots, but now LCCM has no control over it.
I only have one client in the database right now, as
this is a new install of LCCM, and I am testing it out.
The client does not have any profile attached to it. I
do not want to assign it to a profile that may install anew clean OS.
Please tell me what I need to do next.
Answer: LCCM is behaving as designed. If a client is not assigned to a profile, then there is
nothing for LCCM to do with it, so LCCM cannot let it
attach to the server. See the earlier question about how
to assign a client to a profile but leave it intact.
Back to top
Question: Are there any restrictions on the number
of machines that can be defined to the LCCM server.
We will eventually have close to 3000 PC's on our WAN
and we need to know if we will require any secondary servers.
Answer: We would suggest that you split your
machines up over 2 or 3 LCCM servers. Performance
becomes an issue when you have about 1200-1500 clients on a single server.
Back to top
Question: We want to use LCCM in a routed
environment, and using our UNIX
DHCP server. The routers are configured to transmit
BOOTP requests (only to the DHCP server), and subnet
directed broadcasts. All other broadcasts are blocked.
We want to run a remote BIOS update. How can we update
the clients without making an IP-helper entry at the
routers?
Answer: If you want to use LCCM in a routed
environment, you must set a helper-address to the LCCM
Server from every router to which LCCM clients will be
connected. There is NO way around the restriction: the
Proxy DHCP packet MUST be received for PXE to function.
That is part of the specification for PXE. Refer also to
the configuration
requirements for using LCCM in a routed environment.
Back to top
Question: I have created a clone
of a win95 OSR2 system that I want to put onto an
entire classroom of systems. They all have static TCP/IP
addresses. In the network properties, DNS Configuration
screen I can put in the host name via variable LCLI_DNSHOST and LPRO_DNSDOMAIN for the domain. Where
are the variables documented for me to put in the DNS
Server search order and the domain suffix search order?
Answer: This was a bug. It is fixed in LCCM
Service Pack 4. You need to install the service pack and recreate your profiles.
Back to top
Question: I use the
Profile Wizard just to create a framework (Windows
NT® Workstation, SP4) and want to install all the
applications afterwards with my own scripts (I do not
like cloning very much). At the end of the Profile
Wizard created installation (after SP4 is applied), the
LCCM client powers down. How can I arrange a reboot instead?
Answer: You can create a file, LCCM.INI, in
the LCCM installation directory, that contains the following line:
SHUTDOWN=NO
This will prevent LCCM from powering off its clients.
Note that this means that LCCM clients will NEVER power
down: this is an all-or-nothing proposition. You
can of course write your own utility to power down the client whenever you wish.
Back to top
Question: I am having problems getting new
computers to do a PXE remote
boot. I have followed the directions that come with LCCM, but it is not working.
- I am getting a DHCP lease from a Microsoft DHCP server, not located on the LCCM server.
- I chose the IBM DHCP Proxy option when I installed LCCM.
- The network card was setup to boot to the network.
- I am getting a message saying "PXE-E53 No Boot filename received. Loading the boot image
failed."
- I have set the PXEClient option in the Microsoft DHCP server.
- I am not crossing a router.
Answer: In the configuration that you have
described, Option 60 should NOT be set on
the DHCP server. Refer to earlier
question about DHCP setup.
Back to top
Question: I recently experimented with the LCCM
Management Workstation. When I found that you cannot
run the LCCM console on both the LCCM Server and the
Management Workstation at the same time, I decided to
delete LCCM on the Management Workstation.
The effect of this removal was that most of my LCCM
User IDs and groups were deleted from the Windows NT
environment. I could not roll-out any workstations
anymore, because the LCCM clients did not have any
access to LCCM$TMP and LANC$$ anymore.
Answer: To resolve the current problem, you
have to reinstall LCCM on the LCCM server, so that the
necessary groups and accounts may be correctly placed.
To avoid the problem in the future, physically disconnect the Remote Console from the network before
uninstalling LCCM. You will get error messages about not being able to remove accounts, but in this
case, that is actually what you want to happen!
Back to top
Question: Using LCCM v2.51 service pack 3,
patch 1, how can I schedule a client to wake up and boot
from the local Operating System instead of booting off the network.
I would like to be able to do this without having to re-write the alternate boot sequence.
Answer: The only way to do what you want is to
change the Startup boot sequences in the CMOS to
the way you want the client to act. You can do it as
part of the client deployment by using the LCCM CMOS
update facility to change the CMOS as needed. (You could
also do this as a separate Maintenance procedure in LCCM
2.5.1). There is NO way to remotely do it without modifying the boot sequences.
While it is possible to do this on many network
adapters by changing a setting in the network adapters'
configuration, you have to be at the client in order to
do so. Of course, if the client adapter has a DOS-based
configuration utility that is not too resource-hungry,
you might be able to do this as an LCCM maintenance
procedure. This is not possible on any of the currently officially supported network adapters.
Back to top
Question: My customer is a current Token
Ring and LCCM user with 850 desktop users. They are
currently migrating to Ethernet and we want to get LCCM to work on both nets.
LCCM is configured to run with DHCP only. They use
SP6a for Windows NT® and LCCM with SP3 and Patch 1. The
clients are all IBM 300PL's in different flavors.
We tried to put in both an Ethernet and a Token-Ring
card in the LCCM server, but LCCM seems to only use one network adapter.
I have two questions:
- Is it possible to use two network adapters in an LCCM server?
- Why can my Ethernet clients not reach the TFTP server?
Answer: If you wish to use a mixed transport
multi-homed server as the LCCM server, and want to use
it to deploy clients on both segments, you must use the Custom
installation of LCCM and ensure that LCCM is installed
on the Ethernet TCP/IP address, by ensuring that the
correct IP address for the Ethernet card is filled in,
when prompted for the TFTP server IP address during LCCM
installation. In this configuration, the token-ring
clients cannot use PXE 2; you must use PXE 1, which can be downloaded from the IBM
Networking Support site.
Back to top
Question: LCCM is installed on a member
server in a domain. How can I deploy clients into a different domain?
Answer: Install LCCM SP4 and follow the instructions in the readme file.
Back to top
Question: When I scan for my clients, why
won't LCCM wake them up?
Answer: Before or during the scan process, you must turn on the client manually.
LCCM wakes a client by sending a wake-up packet to
the client's media access control (MAC) address. At the
time a scan process is initiated, LCCM has no knowledge
of the MAC address; LCCM discovers the MAC address
during the scan process. Therefore, Wake on LANTM is
not part of the scan function.
LCCM 2.5.1 can wake up a client manually before scan
by inputting the workstation MAC address into the Wake
Clients window of LCCM. This window is accessed through
the Tools selection on the main menu bar.
Please consult the appropriate LANClient Control
Manager Training
and Procedures Guide for more information.
Back to top
Question: LCCM can't see clients on the other
side of my router. Why?
Answer: When a client first attaches to the
network using PXE, it sends out DHCPDISCOVER messages requesting an IP address and a boot file. The
DHCPDISCOVER message is a UDP broadcast message, and as
such, will not normally be forwarded by the router. If
the router does not forward them, LCCM will not see them.
It is possible to configure the router to forward
such messages. On a Cisco router, you would use an
"IP helper-address" configuration
command on the router interface to which a client is
attached. On an IBM router, you would use an "enable
bootp forwarding" configuration command. These
commands instruct the router to forward the messages
related to DHCP to the LCCM server.
Another possibility is that when the IP scope was
specified for a client subnet, a gateway address was not
specified. If a client receives a DHCPOFFER in response
to the DHCPREQUEST, the client needs to know a gateway
address to which it can send the REQUEST message.
To successfully use LCCM
in a routed environment, the following conditions need to be satisfied:
- Subnet Directed-broadcast forwarding must be enabled.
- Proxy ARP forwarding must be enabled.
- The scope for the LCCM client subnet must have a correct router entry for the subnet.
- BOOTP/DHCP forwarding must be enabled. The destination addresses must include the
address of the LCCM server, and at least one DHCP server that serves the LCCM client subnet.
If you are using the IBM DHCP server, this may be the same server. The
destination address may either be the network
address of the servers, or their individual fixed IP addresses.
These are the absolute minimal conditions that
must be satisfied before LCCM will work in a routed
environment. If your network policy will not allow any
of these conditions to be satisfied, then LCCM cannot be used in the environment.
Please check your router documentation for the necessary configuration commands.
Back to top
Question: Why is LCCM performance so very
slow on our network?
Answer: This has usually been observed when parts of the network have been forced to 100Mbps,
full-duplex mode in an attempt to speed network
operations. Please leave all network components to
auto-negotiate their connection parameters. LCCM is a
DOS-based product with no DOS extenders, and so always
works in half-duplex mode. If the network components are
left to take care of it, they often do a great job of
running the network at the best speed, especially if the
network consists of components that are communicating in different modes at different times.
Another reason for slow performance may simply be that your network components are ill-behaved. Some
network cards' default settings are just naturally not
as fast as others', even if they appear to be configured
the same; switches can easily fall into loops, or induce
excessive chatter if not carefully configured
Back to top
Question: We have a LCCM server using RPL
only. The server now has Windows NT® service pack 4
and LCCM Service pack 4. When scanning new clients the
workstation starts to boot from the network and stops with the following error:
Not user level SEC,
RPL.SYS,
Error in server \\LCSERVER
System files not found (non-system disk).
Answer: Verify that the Windows NT Remoteboot Service's Fix Security and Check
Configurations are configured correctly as per the
instructions on pp. 20 of the LCCM 2.5.1 Training
and Procedures Guide.
Back to top
Question: When trying to scan a client I
receive the following error:
RPLDISK.SYS
Error in server \\LCCMPDC
Netbios Addname Failure.
Answer: This is a standard NetBIOS error message, and has NOTHING to do with LCCM, or even the
RemoteBoot Service per se.
When you receive the error "NETBIOS AddName
Failure" in the early stages of starting a
remoteboot workstation, it means that you have given the
workstation a NetBIOS name that is already registered on this segment of the network.
Every computer must ensure that its NetBIOS name is
unique by broadcasting the name. A conflict might occur
because another computer already has this name, or
because a user is logged on under this name.
NOTE: Some network cards can be very persistent about defending their NetBIOS names, even when a
computer crashes.
Back to top
Question: When I try to process my client
after assigning it to a profile, I get the following error messages:
NTSTS Initialization started
NTSTS 0023: A bad parameter value is given for "GATEWAYADDR"
in the NTSTS section of Protocol.ini
NTSTS Initialization failed
IBM DOSLAN Services Driver v5.0 ...
You were logged on but you have not been validated by
a server. Therefore, you may not have permissions to use some network resources ...
NBU0032 Can't open network driver Error 2102
Processing then stops.
How can I get my clients to process correctly?
Answer: The NTSTS driver cannot gracefully
deal with a null string for the gateway address. Your
problem is that you are using a 3rd party DHCP server,
and have a blank router entry. You must provide a router
entry, even if it is invalid. The best thing is to do if
you actually do not have a router on the segment, is to
use the LCCM server as the router address, though any
address will do, even if there is nothing with that
address on the network. Even "0.0.0.0" for the
router option will allow you to successfully process clients in the described environment.
Back to top
Question: At the end of a deployment, I get a
command-line window (on the client) with 2 "Invalid
drive specification" messages in the window.
Answer: The network driver was not correctly
installed on the client. Be sure that you are installing
the correct network driver for the card in the client.
(e.g., have you recently installed a different card, after the client was already scanned?).
To deploy newer machines, download and install LCCM
Service Pack 4. LCCM SP4 wizard-generated profiles will
install the correct drivers. If you use manual profiles,
it is your responsibility to copy the driver files to the correct directories.
Back to top
Question: If I store more than one image
greater than 300MB in the \LCCM\CLNTFILE directory,
I get the following error message sequence on the client:
Bad Command or File Name
Invalid device request writing device C
Abort, Retry, Fail?
After typing 'F' ,
Intermediate file error during pipe
C>
After entering any command on the command line
File allocation table bad, drive C.
How can I avoid this?
Answer: You can avoid this problem by storing
your images in a subdirectory UNDER the \LCCM\CLNTFILE
directory: do not store your images in the \LCCM\CLNTFILE itself.
If you are storing images as zipped files, you can
store more than one image in a subdirectory. However, if
you are storing the image as a collection of files, you
must create a separate subdirectory for each image.
Also, by creating different subdirectories, you will
avoid overwriting files from one image to another.
Back to top
Question: When I try to process a client, I
get the error message: Error
54: Network is busy and processing stops at the client
Answer: This is a misleading message, and
usually indicates a connectivity problem that may have
nothing to do with the network being busy. It has been
observed to happen very frequently when LCCM is being
used in a network that uses switches rather than hubs for connectivity.
You need to disable the Spanning Tree Protocol
on any switch ports that are directly connected to clients.
The Spanning Tree Protocol can remain enabled on any
ports that are connected to other switches or routers.
Spanning Tree Protocol is an inter-switch routing
protocol that is not understood by clients, so there is no penalty to disabling it on client ports.
Back to top
Question: When I try to process a client, I
get the error message: Error 71
and processing stops at the client. If I then log on
manually, as an administrator, I can continue processing.
Answer: This is a Windows networking message,
and means that you do not have sufficient connection
rights to the LCCM server. Naturally, administrators are
not subject to such restrictions; which is why you can connect as an administrator.
It is usually because Windows NT 4 was installed on
the LCCM server by blindly accepting all the defaults.
This means that Windows NT was installed with zero
licenses. Use the Windows NT License Manager on
the LCCM server to set the correct number of licenses
that have been purchased from the license vendor. The
License Manager is located in the Control Panel applet.
Back to top
Question: We will shortly be migrating
machines to Windows 2000 from Windows NT®4.
We have used LCCM before but it does not run under
Windows 2000 server. Is this ever going to be possible?
Answer: You cannot install LCCM 2.5.1 on a
Windows 2000 server. LCCM 2.5.1 can only be installed on
a Windows NT® 4.0 server with Microsoft Service Pack 4
or higher. You can, however, deploy Windows 2000
systems remotely using LCCM 2.5.1 on Windows NT® right now.
Incidentally, if you wish to do remote Windows 2000
deployment using Microsoft tools, you will have to
install Windows 2000 Advanced Server, implement Active
Directory Services, and install the RIS service - no small undertaking!
The net of this is that you can roll out your clients
right now with LCCM as is, so there is really no problem
to solve. Eventually, in the next full version of LCCM, it will be able to run on Windows 2000.
A reference you may find useful is Windows 2000 Deployment, Document ID: SG24-5668-00.
Back to top
Question: Is there any date as to when LCCM will be supported on Windows 2000?
Answer: We are not publishing a date for when LCCM will be able to run on Windows 2000 servers.
Back to top
Question: Can anyone tell me where I can download
the System Installation Tool Kit? The Web site says
it is available from the download site, but I cannot see it there.
Answer: You cannot download the System Installation Tool Kit (SITK). To purchase the SITK
license and CD, please contact your IBM marketing representative.
If you download the LCCM, SMA, and SDA packages from the Web, you do not have a license to use these
products to deploy non-IBM hardware.
Back to top
United States
