Known Issues with the eCS Preview 1

By Jim Burke

When doing a CD boot, after the IBM506S driver finishes displaying the drive configurations, the next few lines crawl along on some machines. (pre 10/7/00)
With these machines, if we turn off the /V option with the IBM506S.ADD driver, the problem goes away. However, we felt that it's important to leave the /V on - at least for the preview version.

Workaround: Relax and have a cup of coffee or something. Go make a phone call.

After the last phase of the installation, upon reboot, you might end up with a black screen and nothing happens after that. (pre 10/7/00)
Workaround: Simply hit Ctrl-Alt-Delete to reboot the computer.

Note: Appears to have been fixed in newer version of MCP (10/29/00)

WarpCenter Display (pre 10/7/00)
When you click on the Disk Space Monitor of the WarpCenter, if you have an unformatted drive, it screws up the WarpCenter display to the right of the disk space monitor.

Work around: remove unformatted drives from the Disk Space Monitor

BTW: WarpCenter still does not display the "Connections" properly when you click on the OS/2 Warp image on the left hand side. The entire network folder does not get displayed (same problem in Warp 4 - looks like it was never fixed).

CRC error with floppy causes a trap D if you ask eCS to retry (pre 10/7/00)
Workaround: Don't select retry. Select stop.

pnpisa.snp hangs certain systems - particularly OEM notebooks (pre 10/7/00)
Workaround: Copy SNDBLAST.SNP on top of PNPISA.SNP. If you're booting from CD, you will have to generate the diskettes and do that to the diskette.

flashpt.snp not found (pre 10/7/00)
Workaround: Ignore the message

Note: This message no longer appears with later drop of the Convenience Pak (10/29/00)

Object Desktop Installation program (pre 10/7/00) and Navigation problem (10/7/00)
You need to read http://www.stardock.com/support/os2/od/odbug20/sec1.htm  before attempting to run the Object Desktop installation program.

The Object Desktop Navigation function does not work properly under eCS (and reportedly has not been working since certain Warp 4 fixpak). We will report this to Stardock and IBM.

Workaround: None

Peer to peer doesn't remember userid password (10/7/00)
This is an old old bug. We will report that to IBM this time.

Workaround: Put in USERID as the userid and PASSWORD as the password. Change them after installation.

Peer to peer service start troubles (10/18/00)
Peer services do NOT start automatically in eCS.

Workaround: Edit the IBMLAN.INI file, the line that now states: wrkservices = ,MESSENGER Change it to read: wrkservices = MESSENGER,PEER Or

You can edit the STARTUP.CMD file, remove the line that reads: NET START REQ and replace it with: NET START PEER

Connections don't start automatically (10/18/00)
The peer to peer CONNECTIONS defined do *NOT* start automatically on LOGON, regardless of what you specify in the CONNECTIONS settings notebook.

Workaround: Include the appropriate NET USE commands in your startup.cmd.

Peer can't see other eCS stations (10/18/00)
Yet another issue is the ability of the eCS machine(s) to see each other (and themselves) on a LAN. This is evident in both the "Create Connections" and the command line "NET VIEW" commands.

Workaround:

Edit the IBMLAN.INI file, and change the line that now reads: srvhidden=yes to: srvhidden=no

DHCPCD.CFG Problem (Prev 1 5/30/2001)
The problem is the dhcpcd.cfg file (in \mptn\etc).

I found it helpful to have the following lines in there: option 1                 # Subnet Mask option 3                 # Router option 6                 # Domain Name Server option 15                # Domain Name option 28                # Broadcast Address option 33                # Static Route option 60 "IBMWARP_V3.1" # Vendor Class option 77 "IBMWARP_V3.1" # User Class If you don't have these, DHCP will not set up DNS and routing automatically.

Suggestion courtesy of Michal Necasek of Scitechsoft.com

Can not install eCS on system with Tekram controller (10/22/00)
The Tekram driver runs in conflict with the sym8xx.add driver that's in the config.sys of the CD boot - causing various strange behavior.

Workaround:

These are the steps you need to follow for a successful install:

You will need to install from diskette. When generating the install diskettes, note the sequence. Makedisk creates disk #1, and #2 first - and then disk #0.

Change the label on the install diskette from VFDISK to DISK 0 by issuing the label command.

Put the Tekram driver TMSCSIW.ADD on your diskette 1. You can eliminate ql510.add to make space.

Modify config.sys (on diskette 1) as follows:

add: basedev = TMSCSIW.ADD set copyfromfloppy=1 rem out: rem basedev = ql510.add rem basedev = sym8xx.add Reboot system with installation diskette #0 in your floppy drive and turn off CD boot

When installing from diskette, got the message Loading, please wait... and goes no further (10/22/00)
Check disk zero's label. If it says VFDISK, change it to read DISK 0 with this command: a:\label DISK 0

Note: Fixed. (10/29/00)

Various WiseMachine errors (10/23/00)
Workaround:

Download the file http://www.touchvoicecorp.com/ecs/wisemachine.zip to your \TVoice directory, and issue the command \TVoice\WiseProg\Unzip\Unzip \TVoice\WiseMachine.zip from the \TVoice directory.

Note: there is a space between Unzip and \TVoice\WiseMachine.zip

I cannot install without the Dani drivers (10/29/00)

Workaround:

After the screen where the slash and backslash are walking over the dots up and down (during the final installation phase from eCS diskette 2) when you see a screen with the text "If you booted from the eCS CD... etc, ... Press F3", press F3 twice to get the command prompt. Then insert eCS diskette one and type: "copy a:\DaniS506.add x:\os2\boot" and "exit" and "Enter" (for reboot).

If the above doesn't work, wait for the machine to reboot and when you see the eCS Prev1 boot blob on the upper left corner of your screen, hit ALT-F1 and then F2 to go to the command line. Perform the above copy and reboot the machine.

The config.sys file at the root of the installation partition x: includes the "basedev=DaniS506.add /a:2 /P:1F0 /a:3 /P:170" line from my eCS diskette 1 already. The installation of eComStation will now continue using the Dani-driver !

3Com ISA 3C509B card causes eCS to trap (10/29/00)
System traps during network install with the following message: c0000005 DOSCALL1.DLL 0002:0000a459

Workaround:

This is a known bug. Most 3Com NIC drivers trap when used with a fast CPU (about 500MHz or more). Go to the 3Com web site and grab the very latest drivers which will solve the problem for *some* NIC models (the 3C509 is among them AFAIK).

==Chuck McKinnis' page NicPak has a good list of drivers and you can get some help there, too. Trap 6 during install (10/25/00)==

During installation got strange TRAP 006 errors.

Workaround:

Turn off second level cache during install.

OS/2 cannot operate your hard drive or floppy diskette (10/29/00)
That message is put out when the country.sys file cannot be read for some reason.. The following is according to Daniela Engert:

Let's assume the hardware and the drivers are working properly (all adapter channels and all attached devices are operated by an eCS driver). Then there are two common scenarios which cause the mentioned error:

1) The bootmanager "sees" partitions differently from OS2DASD.DMD and propagates a bad boot drive letter up to OS2KRNL. In most cases this is caused by unsupported partition types (B, C, and F) located in front of the OS/2 boot partition (in partition scan order). [Note by Kim Cheung: this would happen, for instance, if you have a FAT32 partition in front of your eCS boot partition.   Installing OS/2 with FAT32 around  is an article from VOICE about this.]

2) The adapter channels or units are ordered differently in BIOS and OS/2. This usually happens when more than one adapter is in the system (e.g. an on-board IDE controller plus a SCSI host adapter or an additional EIDE controller, either on-board or a PCI plug-in, more than one of them equipped with DASD units). In this case, their order of appearance in CONFIG.SYS is important! Most BIOSes have options to reorder the adapter channels, and the OS/2 drivers are required to match that (either by reordering the ADDs in CONFIG.SYS or - in rare cases - by some ADD options trickery). Please note: this is not an absolute requirement, but it is hard to get away with a non-matching adapter/unit order.

Unfortunately resolving such a problem requires major knowledge and can't be done automagically.

DFSEE and partitioning problems
To see what partitioning information you have on your drives, get Jan van Wijk's program, DFS.EXE from here: http://www.fsys.demon.nl/dfsee.htm It is almost an essential tool for some of the problems you might encounter.

LVM can't access old boot manager partition (10/29/00)
There have been report that LVM fails to work with old boot manager partition left there from previous versions of OS/2. However, others had indicated that they have no problem.

Workaround:

If it affects you, remove the old boot manager and reinstall it under LVM. Reading material: LVM, FDISK and Partition Magic ; A Short Introduction to LVM and JFS (both from VOICE)

TRAP E from In-Joy and DIOP dialer (pre 10/31/00)
{Could be trap D, too.}  TRAP 000e ERRCD=0002 ERACC=****  ERLIM=******** EAX=3b81939c EBA=f7161c14  ECX=00000000  EDXX=3b81939c ESI=f7161c14 EDI=00000000  EBP=00004ad0  FLG=000122246 CS:EIP=0168:f77aa125d CSACC=c09b  CSLIM=ffffffff SS:ESP=0030:00004ab0 SSACC=1097  SSLIM=0000401f DS=0160 DSACC=c093  DSLIM=ffffffff  CRO=80010013 ES=0160  ESACC=c093  ESLIM=ffffffff  CR2=3b81930a0 FFS=0000 FSACC ****  FSLIM=******** GS=0000 GSACC=****  GSLIM=********

The system detected an internal processing error at location
 * 1) 0168:fff1f322 - 000e:c322. 60000, 9084

07860652 Internal revision 14.062_W4 

Workaround:

Try the new stack fix from ftp.service.boulder.ibm.com/ps/products/tcpip/fixes/v4.3os2 and get the file dated Jan 29 10:26, called  ic27649.exe, and which is  844,651 bytes or you can get it from the In-Joy  web site.

You can also try the WSeB UNI kernel at ftp://testcase.boulder.ibm.com/ps/fromibm/os2/. This is for single processor machines. Use the smp kernels for multi-processor ones. Don't use the W4 kernels. They are for Warp 4 based machines and eCS is based on the warp server for e-business kernel.

NOTE: Try these kernels after you try the stack fix and ONLY if you are experiencing trouble.

Newer JFS version (pre 10/31/00)
There is a newer JFS version we have been told that ftp://testcase.boulder.ibm.com/ps/fromibm/os2/jfs1009.zip is a good fix to have and ftp://testcase.boulder.ibm.com/ps/fromibm/os2/jfs_miss.zip has some missing jfs trace files.

I can't install eCS past 1024 cylinders (pre 10/31/00)
Depends on your hardware. If you hae a new bios that understands Int13 extensions, then you can install past the first 1024 cylinders. Older Pentiums don't have this, BTW.

Depending on your hardware, this can mean 512 MB, 1 GB, or 8 GB. With Logical Block Address (LBA), you can boot from up to 8 GB. This includes most Pentium class systems. With 486s and earlier, you're likely to be restricted to 512 MB or 1 GB, depending. [Note that this isn't a CPU restriction; it's a restriction due to the BIOS. For example, a new BIOS for a mainboard may allow it to boot from < 8 GB to anywhere on a large hard drive.]

Can't install eCS to certain partitions (11/03/00)
I created a partition on my secondary master IDE drive and tried to install eCS to that partition. The installation seems to go perfectly until I remove the diskette and reboot, I get the message, "operating system missing, system halted." If I reboot I get my boot manager choices from before the installation attempt. It is as though LVM changes made in the install don't "stick."

Workaround:

You will need to replace the Warp 4 boot manager with the one created by LVM. LVMGUI has a menu item which says Bootmanager - use that. Click here for an image.

1608 error during network install (11/03/00)
This is a long standing issue dating as far back as Warp 3. Some NIC cards don't like to be poked by the MPTS sniffer. Warp Server install used to have a /NS option to deal with this but that option is not available with eCS.

Workaround:

Unplug the NIC card from your computer during install and reinsert it afterwords.

SmartSuite Install error: " Noncurrent OS Detection " Make sure 4os2 isn't used in config.sys.

Unpack error when installing SmartSuite 1.5 (11/29/00)
> >> EPFIE196: An error occurred while executing M:\INSTTEMP.EPF\EPFIUPK2.EXE O:\ss\english\WORDPRO\LWPWK6O.DL_ L:\LOTUSW4\WORDPRO.

Workaround:
 * 1) Download http://service.indelible-blue.com/eCSbeta/LWPWK6O.EXE (This link may not work anymore.)
 * 2) run LWPWK6O.EXE to extract LWPWK6O.DL_
 * 3) xcopy the smartsuite directory to a harddrive example: XCOPY G:\SS C:\SS\ /S/E/V
 * 4) Copy the extraced LWPWK6O.DL_ file into the \ss\english\wordpro directory.
 * 5) From the \ss directory, run INSTALL.CMD

SYS3175 when installing SmartSuite 1.5 (pre 11/29/00)
Installation stopped.

Workaround:

(a) Download the updated WiseMachine package at http://www.touchvoicecorp.com/ecs/wisemachine.zip

(b) Make sure WiseMachine is not running.

(c) From the \TVoice directory, unzip wisemachine.zip (with the overwrite and directory options if your unzip doesn't do it automatically). You get a new \TVoice\WiseProg\WiseManager\WiseMachineLocal.exe, \TVoice\WiseProg\WiseManager\SmartSuite\*.*, and \TVoice\WiseProg\DeskTop\CreateSmartSuite.cmd.

(d) Start up WiseMachine, Drag-n-drop SmartSuite to your destination drive. Modify the destination top-directory to say lotusw4

(e) Check the box "Use Unpack..." and click Yes at the warning message screen

(f) Insert the Lotus CD into your CD-ROM and point Source to one of the files in the \english directory of the CD (or whatever language directory there is).

(g) Click Install

This will unpack and transfer the ENTIRE Lotus CD to your hard drive.

(h) Close down WiseMachine.

(i) DON'T REBOOT THE MACHINE YET. Run the Lotus supplied install program. For my case, the install program doesn't fail with SYS3175 any more.

I wasn't able to get rid of this step because I think Lotus is putting some registration and path information into the emulated Windows registry files and I don't have time to investigate more details there to be able to get around it with WiseMachine. I should be able to do so in the future.

(j) Restart WiseMachine again, go to Installed Applications, right click on SmartSuite and do a Create ICON.

Using this method, you should be able to get all but Organizer to work.

Smart Suite Launch Problem
Error opening ole

>It looks to me as if you have to physically delete those BASEDEV lines, >rather than just REM them out. >... >Peter

Thanks for your help ! I deleted the unwanted lines from config.sys and it works now.

Removable Media Problems
because of broken removable media support (no matter if ZIP, ORB, or PCMCIA attached PRM devices). Almost any time I inserted/replaced media I got either or no matter if FAT or HPFS. I couldn't find any reproducable pattern.
 * a TRAP 3 in OS2DASD or OS2LVM (I don't know which one)
 * no reaction on "refresh removable media"

There is a new driver available for these removable drives.

There are two new OS/2 device drivers now available on the OS/2 DDPak (http://service.software.ibm.com/os2ddpak/) that may be of interest to some of you. They're currently beta, but will become "IBM Supported" when FVT completes.
 * Driver for the internal 56K modem on the ThinkPad T20 family. It's listed under MODEMS/IBM.  Yes, I know this is a WinModem, but it does work under OS/2 with this driver.
 * Driver for Parallel Port-attached ZIP drives, ZIP 100, ZIP Plus, and ZIP 250. It's listed under REMOVABLE DISKS/IOMEGA.

icmp: unknown protocol error
When you run ping or ftp and other programs you get an ICMP* error. From prev 2 and above, some files are missing from the installation cd. In this case the solution is to use another version of Warp or eCS and copy these files to the new installation mptn\etc directory. x:\mptn\etc\protocol 6K x:\mptn\etc\services 213K (missing but doesn't seem to be necessary to fix this)
 * Internet Control Message Protocol

Dos Lan Bug
Problem has been identified. It is an INSTALL problem (MPTS/TCPIP) where CONFIG.SYS is not updated. Actually, the MPTSCFG.INF file (VIEW it) had the answer as well. See the LAN Virtual DD section.

Two lines are not inserted into CONFIG.SYS during Install, the required Virtual drivers for DOS. DEVICE=drive:\IBMCOM\PROTOCOL\LANVDD.OS2 DEVICE=drive:\IBMCOM\PROTOCOL\LANPDD.OS2

SIO2K.SYS will not load
"Empty device table (no devices defined). SIO2K.SYS will not load."

CONFIG.SYS contained, very close to the end of the file: device=c:\sio2k\sio2k.sys logfile=c:\sio2k\sio2k.log device=c:\sio2k\uart.sys logfile=c:\sio2k\sio2k.log device=c:\sio2k\vsio2k.sys logfile=c:\sio2k\vsio2k.log vIrqList(3,4)

By changing the order to: DEVICE=E:\SIO2K\UART.SYS LOGFILE=E:\SIO2K\SIO2K.LOG DEVICE=E:\SIO2K\VSIO2K.SYS LOGFILE=E:\SIO2K\SIO2K.LOG DEVICE=E:\SIO2K\SIO2K.SYS LOGFILE=E:\SIO2K\SIO2K.LOG the problem was fixed.

NB: This has nothing to do with eCS unless you are planning on using the Ray Gwinn SIO drivers.

In WiseMachine, when you first click on "Installed Applications", there is nothing under the "Applications" branch. (pre 10/7/00)
Workaround: Right click on the "Applications" folder and do a refresh.

Note: Fixed (10/29/00)

Not every Rexx script under WiseMachine has been fully tested. (pre 10/7/00)
When you do a drag-n-drop deployment of an application, WiseMachine executes a Rexx script for that application. These scripts were adopted from the network version and have not been fully tested under a local non-diskless station.

Workaround: If you have problem with certain application, be sure to go to our web site at www.ecomstation.com  and follow the link for more recent instructions.

Note: Over 2/3 have been tested (10/29/00)

WiseMachine is not WPS aware (pre 10/7/00)
The preview version of WiseMachine is not a SOM application. What that means is that if you moved an installed application from its original location to a new location, WiseMachine is not aware of that automatically.

Work around: What you need to do is to bring up WiseMachine again, do a drag-n-drop to the new location, and uncheck the "Copy program to destination location" option, and do an Install. This would force WiseMachine to update all of its record of that application. In the future, WiseMachine will be updated to include the ability to do this automatically.