• Welcome to OS2World OLD-STATIC-BACKUP Forum.
 

News:

This is an old OS2World backup forum for reference only. IT IS READ ONLY!!!

If you need help with OS/2 - eComStation visit http://www.os2world.com/forum

Main Menu

Help - eCS fails to boot

Started by Pete, 2010.07.10, 19:23:08

Previous topic - Next topic

Pete

Hi All

Warning: Long Post!

I have bumped into a problem that I cannot resolve: I cannot boot either of my eCS installations and nor can I boot

from eCS2.0GA or RC7 install CDs.

I can however boot my WindowsXP installation - so I'm posting this from WinXP using IE  :-(

I guess I should start by listing my system hardware and disk Volumes.

Hardware: Asus M2N mainboard with AMD 4400+ dualcore cpu, 2Gb RAM, ATI X800 graphics, 1 * Sata2 160Gb disk, 1 *

Sata2 80 Gb disk (both Samsung), 1 * LG GSA-4167B burner, 1 * LG GDR8164B dvd rom, 1 * floppy drive, 1 * Internal

USB 4 slot card reader.

Disk Volume layout,
Disk 1: C: WinXP NTFS, IBM Boot Manager, D: WinApps FAT32, E: DOSApps FAT, F: DOSApps FAT, G: Downloads JFS, H:

eCS2.0GA BootableJFS, I: UserData JFS
Disk 2: J: eCS(OS/2)Apps JFS, K: MoreApps JFS, L: eCS2.0RC6a BootableJFS


Initial problem: I was booted from eCS2.0GA on H: and running WinXP in VPC - VPC is on J: - when the system simply

stopped working; no response to mouse or keyboard, required a press of the system box Reset button.


The reboot got as far as loading the Desktop but stopped working when starting File and Print.

I used the Reset button again and booting stopped before the "loading unicode ..." message appeared; The next

attempt stopped at the same place.

I then tried booting my RC6a installation on L: as that does a full chkdsk during boot. The chkdsk went fine but the

system stopped booting with the loaded USBRESMG.SYS message onscreen.

I attempted to boot from L: again and the system stopped booting in the same place.


I have attempted to boot to a command line, using Alt-F1 followed by F2, on both eCS installations but boot never

completes.

I then attempted to boot from eCS2.0GA and RC7 install CDs. Neither of those get as far as the install gui -

repeatedly.


I then opened the system box and cleaned out the dust - this can sometimes cause weird problems. However, it made no

difference, I still could not boot to either eCS installation nor could I boot from either eCS Install CD.


Just because I had not yet tried it I decided to boot the WinXP installation. No problems, XP booted fine and worked

without problems.

This suggests the problem is *not* hardware - in my limited experience Windows often "screams" first when hardware

is a little flaky.

I had been wondering if my graphics card was the cause of the problem as eCS boots fail close to switching to

graphics mode to display the Desktop - but booting and running XP now makes me think not.


I then had the following thought: WinXP on C: does not do anything with disk 2 ie J: K: and L: volumes and both eCS

installations do - so maybe that disk is failing.

I powered off, disconnected disk 2 and attempted to boot from eCS2.0GA on H: That failed the same as previously.

With disk 2 still disconnected I attempted to boot from the eCS Install CDs. Neither of those could boot to the

installer gui still.

Looks like the problem is not disk 2.


Leaving disk 2 disconnected I then disconnected disk 1 and attempted to boot from eCS install CDs. Failed as before.

At that point I gave up for the night and went to bed - it was about 3am.


Today I got back onto this problem at around 1pm.

I reconnected both drives and attempted to boot eCS2.0GA on H:

Boot was successful and the Desktop appeared, I logged onto the network after F & P had started.

I opened h:\os2\boot\config.x in the text editor with the intention of cleaning it up - removing all the unwanted

DOS lines that the eCS2.0GA install had insisted on putting in that file despite the deselection of DOS/Win16

support.

Shortly after opening the file the system locked solid and required a press of the Reset button.

I attempted to reboot from H: which failed the same as the previous day - as did an attempt to boot from L:

I also tried booting from a BootableCD created about 2 years ago - the hardware is more or less the same - but that

failed to boot in a similar manner to the eCS Install CDs; it simply stopped.


So, here I am running WinXP from C: and wondering why this works when 2 usually healthy eCS installations do not

boot - and if H: does boot it locks solid shortly after.

I found a BootableCD that I created about 2 years ago. As the hardware has not changed dramatically I think it

should have booted - but it failed to get to a Desktop.

I doubt there is a hardware problem involved...


I have downloaded and used the DFSee bootableCD but could spot nothing obviously wrong with volumes/partitions -

probably the interesting thing is that this CD did boot OK.


I am wondering if there is some sort of virus/rootkit problem though.

My reason for thinking along those lines is that none of the eCS boot options - hard disk and CD - available to me

can boot but WindowsXP and the DFSee CD can.

When booted from either eCS installation there is no access to WindowsXP on C: - I do not even use the ntfs driver

with read only access to ntfs volumes.

Could my eCS system have caught a bug? - and if "Yes" How do I get rid of it?

Any ideas on the cause and solution to my eCS non-booting problem would be appreciated.


Regards


Pete

RobertM

Do you have any drives (or printers with MSD ports or devices in them) plugged in via USB? If so, you may wish to unplug them and try again. I have an XP machine (also an Asus motherboard) that has boot problems because the BIOS will re-assign boot device priority when it detects one on bootup.

During booting to XP, it will sometimes hang (or it will just sit on a black screen and not boot at all) until I unplug the device and restart (via reset or power off/on) the computer. I suspect I'd have similar issues booting eCS if I had eCS on this machine.

Best,
Robert


|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


Pete

Hi Robert

Thanks for the thought but it does not seem to have any bearing on the problem here.

I'm currently posting from my fixed eCS2.0GA system - he said hopefully as the system has now been running for 3 hours without problems :-)

I'm not really sure what the problem was but it seems to be graphics card related.

Both eCS2.0GA and RC7 installations use SNAP.

When attempting to boot 2.0GA or RC7 Install CDs I selected SNAP as the graphics driver because I knew it *should* work.

I did attempt to boot the 2.0GA Install CD using Panorama but that did not get as far into the boot process as the previous boot attempts using SNAP before it hung.

Today I tried to boot from 2.0GA Install CD using GRADD - and it worked!

I booted to the Maintenance Console and ran chkdsk against all drives. This does give me a little concern as it reported the following for drive i:

LVM reports 1 bad block ... 0 transferred to JFS bad block list.


I would have expected a bad block to get transferred...

Anyway, having completed all chkdsks required and tested that I could access files on all drives without problems and had a check of the config.sys files for both eCS installations I rebooted to eCS2.0GA on hard drive.

The system booted fine and, as mentioned previously, has been working fine for (over) 3 hours.


As the problem occurred when the system crashed while running IE/WinXP in VPC/2 I am wondering if some sort of nasty file got downloaded in the background causing SNAP to fail to initialise the graphics card properly on reboot leading to a hung boot - or system crash within minutes if the boot got to the Desktop.

Since I posted this problem I have not played with the system box internals - all I have done is boot the 2.0GA CD using GRADD as the graphics driver and that seems to have resolved the problem(s).

Regards

Pete







RobertM

Quote from: Pete on 2010.07.11, 18:46:30
I booted to the Maintenance Console and ran chkdsk against all drives. This does give me a little concern as it reported the following for drive i:

LVM reports 1 bad block ... 0 transferred to JFS bad block list.


I would have expected a bad block to get transferred...


I've never seen that on a drive that hasnt been booted from (ie: chkdsk of a drive when booted from a CD). I am wondering if the block was already transferred on a previous chkdsk?


|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


ivan

Hi Pete,

It could just be as you say, something got put on your boot drive that blocked the graphics card initialization.  Since you were using vBox your eCS boot drive would be the equivalent of C drive on windows as far as most malware is concerned.

I recently had to clean up a friends windows machine that had one of those fake AV bugs - that one took over the video system so it could display the fake warning messages.  In their case it was also an ATI graphics card.

By using the GRADD driver you bypassed most, maybe all, the advanced functions using the fallback VESA mode which appears to run even if there is malware trying to do nasty things.

Now you are up and running I would suggest you run chkdsk again from a command prompt and also check your ini files for anything strange.

ivan  

Pete

Hi Ivan

No, it could not be anything put on a boot drive that prevented graphics card initialisation - I tried booting both eCS2.0GA and RC7 Install CDs with both hard disks disconnected and boot failed.

I was not using VBox - I was using VirtualPC/2.

Would a virus see any boot drive when running Windows with either of the above though? In both cases the host system drives are accessed via networking.

Nothing strange found in INI files.

I set config.sys to do "proper" chkdsk at boot and do not see the "LVM reports 1 bad block ... 0 transferred to JFS bad block list" message that I got when running chkdsk from eCS2.0GA Install CD.

A final question: Why did GRADD work but Panorama did not? - I believe Panorama is a VESA driver.

I do not seem to be able to find anything about any Windows malware/virus that targets graphics cards so maybe the problem was some sort of glitch somewhere within WindowsXP/VPC/SNAP rather than a Windows "nasty".


Regards

Pete



klipp

#6
Hi Pete , I have what appears to be the same exact thing happening! I've been working on this box all day alternatly using a test partition with 2.0 GA, also another partition with RC-6 and a third partition with 2.0 GA for production, all giving the same intermittant results you posted.(and all use SNAP)! The last thing I did before this all started (about 4:00 P.M.) was to boot up to Windoz XP and do some printing on a new printer I bought that I can't seem to get to work with eCS. And yes, now I can boot to Windoz but no more eCS. I can also boot to Linux.
  Strange, I think my box has caught a "Social Disease"!
  I'm going to give it a rest and try again tommorrow. Fortunatly I have other machines on the LAN I can still use.

klipp
********************************************************************
Windoz Free! eCS 2.0 serving Linux, Windoz & OS2
rklipp@megared.net.mx
********************************************************************

Pete

Hi klipp

Interesting to know that it is not only my system.

The only advice I can offer is to try booting from eCS2.0GA Install CD, go into the preboot menu and select the GRADD video driver as that seemed to resolve the problem here. Hopefully that will work for you.

As to the cause of the problem I still have no idea. I do wonder if it is related to nVidia and ATI suggesting using graphics processors for general programming tasks and providing toolkits to achieve this. It may not even be a "windows nasty" causing the problem but badly written "genuine" code trying to utilise the GPU.


Is your new printer not supported by CUPS? - not much chance of it being supported by the OS/2 OMNI driver.


Regards

Pete

klipp

Well Pete, Seems giving it a good night's rest seemed to help, I'm using it now. Possibly NVidia related as you mentioned or a heat problem as I now have the covers off the box. I'm sure it'll be back, next thing I guess I'll try is a different Display Adapter. I'll let you know if I get any smarter about this problem.
  As far as the Printer is concerned, its a Samsung CLP-310, and it does work with CUPS. Prints a "test page" just fine, in color! The problem being with the "pin.exe & ppdenc" software that is supposed to convert the working Foo2qpdl.ppd file into something eCS can understand and create a working printer desktop object so that programs like Open Office can print to the printer. I've been screwing with that for over a week now and have gotten nowhere. Supposedly its backwards compatable to the Samsung CLP-300 which Paul has working but that doesn't seem to be true either.
  Thanks for the feedback.

klipp
********************************************************************
Windoz Free! eCS 2.0 serving Linux, Windoz & OS2
rklipp@megared.net.mx
********************************************************************

Pete

Hi klipp

If it is a case of pin.exe adds the printer to the ecups.drv OK but there are still problems then you may want to try a version of the ecups driver that does not have any ppd files already added. I understand that there is some sort of limit on the number of ppd files that can be added into the IBM postscript driver and when that limit is reached problems occur.

Paul did a "naked" (no ppd files included) recently ... just trying to find the url in the ecups ng ... Aha! From Pauls post in the ng, thread: "Kyocera-Mita FS C5100DN will not import into ecups.drv":-


Please try using http://smedley.info/ecups-driver-20100705.zip to
import into.

This is the same code as the March 2009 release, but has NO other PPD
files imported into it.

There are some issues in the IBM code where the driver gets 'maxxed
out' if too many PPD files are imported - both the march 09 ecups.drv
and the 2010 ecups-hp are near this limit.

I'm thinking that in future, some kind of 'wizard' that scans
\cups\etc\cups\ppd and looks for the actual PPD files that are
installed by the user, and imports these into ecups.drv is a better
solution than trying to supply driver packs with 100's of ppd files
already included.


Hope the above is of help

Pete



klipp

Hi Pete, thanks for the idea. I tried using "Naked version" but when I went to use pin.exe I get a "reading driver extended attributes fail"-- the system cannot open the device or file specified, error. I tried using eautil but also with no success, possibly I'm going about using it wrong. I don't really understand how eautil is supposed to work. Should I be using it on the .PPD I'm trying to convert?

Thanks again for all the help.

klipp
********************************************************************
Windoz Free! eCS 2.0 serving Linux, Windoz & OS2
rklipp@megared.net.mx
********************************************************************

Pete

Hi klipp

Yes, sounds like you need to use eautil to join the ecups.ea file containing EAs to the ecups.drv file.

Open a command window in \OS2\DLL\ECUPS and enter this command:-

eautil ecups.drv ecups.ea /J


That should work - and the ecups.ea file will disappear.

You should then be able to add the PPD file to ecups.drv using pin.exe without problems.

Regards

Pete

klipp

Close, but no cigar, Pete! using eautil with the /J option gives the error SYS1829: The extended attribute (EA) join operation is not allowed. So, looking up SYS1829 it says to use the /O option (which appeared to work) the ecups.ea was no longer in the directory. Then using pin PPD I get:
Postscript resourse packaging utility version 1.00
Copyright blagh blagh

Producing Printer Device PAK
reading driver extended attributes
  fail
The system cannot open the device or file specified.

fail

Looks like it doesn't like pscript.drv!! Another thing Pete, is the foo2qpdl.ppd file is 9 pages long! I've tried condensing it by elimination all the REM's (*%)'s and deleting everything except the default entries but that didn't work either.

klipp
********************************************************************
Windoz Free! eCS 2.0 serving Linux, Windoz & OS2
rklipp@megared.net.mx
********************************************************************

Pete

Hi klipp

SYS1829 means that the EAs are already attached - usually.

Is the "fail" related to the EAs or to the attempt to add the PPD files to ecups.drv?

If it is the PPD file then I would guess that the foo2qpdl.ppd file is too big.

It may be worth posting in the cups ng news://news.gmane.org:119/gmane.org.netlabs.ecups.devel in the hope that Paul may be able to look into the problem.

Sorry, I don't have any other suggestions at the moment...


Regards

Pete



klipp

Thanks Pete, Paul is aware of the problem and duplicated it and said he will look into it when he gets time (see my post  "Samsung CLP-310 - need driver!".
  Thanks again for the help.

klipp
********************************************************************
Windoz Free! eCS 2.0 serving Linux, Windoz & OS2
rklipp@megared.net.mx
********************************************************************