Troubleshooting iChain 2.0 Installation Problems
Articles and Tips: article
Novell, Inc.
01 Apr 2002
Installing iChain 2.0's Proxy Server is a straight- forward operation. You simply insert the iChain Proxy Server CD, boot the server, type "BLAST <Enter>" and accept the license agreement! However, now that you can install the iChain 2.0 Proxy Server on different hardware platforms (unlike its v1.5 counterpart that required dedicated hardware), customers have seen an increase in problems during the installation process.
The goal of this Network Troubleshooter column is to identify the various stages that the iChain Proxy Server installation program goes through, point out potential problems in each stage, and give troubleshooting tips to help you overcome these problems.
Note: You should perform a low-level format on the server's disk drives prior to installing the iChain Proxy Server.
The DOS Image Stage
The whole process begins when the server boots from the iChain CD, which loads DOS and allows you to run a program called BLAST.
BLAST creates a DOS partition on the first server hard disk drive and copies binary-image data sector-by- sector from the CD onto the hard disk, including the "bones" of the NetWare server. Other data includes files that are used in subsequent stages, including the master boot record, the DOS partition, and the NetWare .NCF files that are copied to the \NWSERVER directory. This stage ends by rebooting the server.
Troubleshooting the DOS Image Stage
Typically you see very few problems at this stage. If you do have trouble after running the BLAST program, check to see whether you can access the server's hard disk drive after rebooting with bootable floppy. If you cannot, there's an issue with the disk or controller. Check to see what type of disk and controller are installed in the server and replace them with different ones if the installation cannot proceed beyond this stage.
The Hardware CDT Stage
The Customizable Directory Tree (CDT) stage begins as the server reboots from the CD. In the previous stage, a special file named RDW was written to the server's disk. The presence of this file tells the system to transfer DOS execution to the disk drive.
The NetWare server then runs the STARTUP.NCF and IMAGE.NCF command files. As the program detects the hardware (by running HDETECT.NLM), it copies any needed device drivers from the CD to the local hard disk. HDETECT.NLM writes the STARTUP.NCF and RESPONSE.SAV files to the disk as it performs the hardware detection process.
During this stage, the IMAGE.NCF file is modified for use in the next stage. The ouput from the START- UP.NCF and RESPONSE.SAV files that were created through the hardware detection step is deleted, as the original STARTUP.NCF will be used for the next stage. This stage ends by restarting the server.
Troubleshooting the Hardware CDT Stage
As with the previous stage, very few problems are seen at this stage. Any problems at this point are likely to be related to problems that existed in the previous stage, or result from a bad CD or CD-ROM drive. If you have performed all the troubleshooting steps from the previous stage and still have problems, try using a different iChain CD.
The NetWare Image Stage
When the server comes up, it runs in hardware detect mode and loads any needed drivers. The modified version of STARTUP.NCF and a new RESOURCE.SAV are written to the server's disk drive.
The NWIMAGE.NLM program then runs. It first creates a NetWare partition large enough to hold the SYS and LOG volumes. Then the binary image of the SYS volume is copied sector-by-sector from the iChain CD to the disk. At this time, any needed commands to load drivers (as well as minimum and maximum packet receive buffer commands) are added to the STARTUP.NCF file. NWIMAGE.NLM then creates the LOG volume on the server's disk drive.
The OEMINST.NLM program runs to copy any Customizable Directory Tree (CDT) components from the CD to the SYS volume. The IMAGE.NCF file is also modified for the next stage. The DETECT.NCF file is deleted to avoid running the hardware detection process in the next stage. A command is issued to restart the server in preparation for the next stage.
Troubleshooting the NetWare Image Stage
Problems are most common in this stage of the installation. The typical symptom will be that the iChain console screen hangs, showing no progress whatsoever. Problems at this stage tend to all be related to hardware. To get around the problems, you can perform the three major troubleshooting steps described below.
Disable All Advanced BIOS Settings at the Hardware Level
During bootup, select the option to bring up the system BIOS information (for example, on Dell computers press the <F2> key).
Disable any RAID technology, as iChain servers gain nothing from RAID and the product was not designed to use it.
Enable a single CPU. This can be done by selecting CPU Information and making sure that, if there are two CPUs, you remove one as an enabled selection.
Select the Integrated Devices option and disable all on-board NICs. This step may not be required, but some systems ship with LAN cards for which iChain doesn't have a driver (such as cards which use the BE27.LAN driver).
Disable any advanced CPU caching features, such as Turbo CPU mode.
You do not need to disable the serial ports.
If you haven't done so already, perform a low- level format on the server's disk drives prior to installing the iChain Proxy Server.
Reboot the iChain Proxy server hardware again and re-enter the BIOS settings. Make sure that all the changes you made are registered correctly.
Reboot the server with the iChain CD and try running the install program again.
Determine Whether Drivers Are the Problem
If you have problems with the new install after you have disabled the advanced BIOS options described above, go into the debugger by pressing <Alt>+<Left-Shift>+<Right-Shift>+<Esc>). Type "v <Enter>" to initially toggle to the console screen (you then simply press the v key to toggle between screens).
Scan through all active server screens at the time of the hang.
Check to see if the system console shows that it's in the process of loading the CD drivers, hard disk drivers, LAN drivers, or that it is trying to locate a response file. If this is the case, it is likely that there is an issue with the default driver that the iChain install process loaded.
Note: Pressing <Alt>+<Esc>, which is the usual way to toggle between screens, will not work while the server is in debugger mode. Note also that the server is effectively halted while in debugger mode.
Many of the hard disks or LAN controllers detected have more than one associated driver. The installation process will stop at this point, waiting for the user to select the driver. However, since iChain locks the server screens during install, you cannot make the selection to continue the install process.
If there is an active screen waiting for user input, write down the drivers it lists.
Locate the correct driver for the peripheral being loaded and restart the installation using the steps shown under the next heading.
Updating Drivers During the iChain 2.0 Install
The steps below outline how to update IDE and SCSI drivers. The CD-ROM driver will be the most common failure on an install. Before attempting the driver update, try swapping the CD-ROM drive with one from a different manufacturer, or try using an external SCSI CD-ROM with the correct driver.
Install the proxy server via the iChain CD.
After the reboot that occurs just after the DOS Image Stage (after it creates the DOS partition and copies the drivers and startup files to the \NWSERVER and \NWSERVER\DRIVERS directories), press <F5> or <F8> as DOS boots. The <F5> key prevents the startup files as well as COMMAND.COM from loading, while the <F8> key gives you a line-by-line prompt through the startup files.
Load COMMAND.COM.
First try copying only the IDEATA.HAM and IDECD.CDM files to both the D:\NWSERVER and D:\NWSERVER\DRIVERS directories for safe measure. (Drive letter C will likely be the CD-ROM drive letter designation.)
Reboot the system and let the installation proceed as planned.
If the installation continues to fail after this, the failure will probably be while it is creating the LOG volume on the system. If it hangs, go into the debugger and use the "v" debugger command to switch to the system console (the <Enter> key will toggle you through the screens while in debug mode). Check to see whether or not the server is going through the volume creation step--you should see a message on the system console indicating that the volume is being created.
If the volume creation does fail or hang, start the installation over as described under "Trouble- shooting the Hardware CDT Stage". Copy updated SCSI and IDE adapter and hard disk drivers--ADPT160M.HAM and IDEATA.HAM, as well as SCSIHD.CDM and IDECD.CDM (on a Dell system)--to the \NWSERVER and \NWSERVER\DRIVERS directories, respectively.
The NetWare/COS Partitioning Stage
At this point, the server comes up and runs the commands in the IMAGE.NCF file. It also runs the AUTOVOL.NLM program to mirror the NetWare partitions to a secondary disk. The DOS partition is cloned and a snapshot of the contents of the SYS volume is copied to the mirror disk. However, the LOG volume is not copied.
To finish preparing the disk drive(s), this stage creates the Cache Object Store (COS) partitions on all disks. This is followed by a hard reboot of the server.
Note: Where only one disk is present on the system, no cloning/mirroring is performed. In this case, the installation process moves immediately to the Completion stage without rebooting or restarting the server.
The Completion Stage
In the Completion stage, the server reboots from the CD and because the RDW file is present, it transfers OS execution to the hard drive. From the hard drive, the server is brought all the way up into an operational state.
As the server comes up, it runs AUTOVOL, which in turn runs the AVLOAD.NCF file. This causes the RESPONSE.SAV file containing the output from hardware detection stage to be parsed and to look for LAN cards. RESPONSE.SAV modifies the server configuration using INETCFG.NLM APIs to add the LAN hardware to the server configuration.
Because of this, it is recommended that you shut down the server, remove the iChain CD, and restart the server so that the iChain server will run completely from the hard drive.
Troubleshooting the Completion Stage
Make sure that the LAN drivers loaded through the NETINFO.CFG file correspond to the drivers you expect. Many LAN cards use the Intel N100 chipset, and each of these cards can run with mutiple LAN drivers. Some drivers (particularly newer ones) provide better performance than others.
For example, if the older CE100B.LAN driver is enabled instead of the newer N100.LAN driver that should be enabled, it could result in an increase of HTTP 504 errors under heavy load.
* Originally published in Novell AppNotes
Disclaimer
The origin of this information may be internal or external to Novell. While Novell makes all reasonable efforts to verify this information, Novell does not make explicit or implied claims to its validity.