[kDn]

doveman, on Nov 11 2009, 09:44 AM, said:

Has anyone had any success testing this in a Virtual machine?

Yes, I tested this method of install on VMWare many times.

[doveman]

Oh well, at least now I know it's something I'm doing wrong!

I've been doing some testing with XP install files on the HD and it seems the errors I'm getting in Vmware are caused by an incorrect "SetupSourceDevice = " line in TXTSETUP.sif.

So, what should I set this to before making the ISO?

This post has been edited by doveman: 11 November 2009 - 05:12 AM

[cdob]

travis_san, on Nov 10 2009, 02:21 PM, said:

Yes, I did RAM load this way by linking the usb drive as a virtual drive.

Is this how you tested it in a vm or did you use a different method?

As far as I know VirtualBox mount the USB stick to a virtual internal hard disk.
There are no USB drivers used. This does not test USB boot.
Installation way work anyway, but test internal hard disk behaviour.

I'm using VMWare Player, mount the USB drive, boot a PLoP Boot Manager floppy image and continue USB boot.
Windows setup use USB drivers for the USB drive. This does test USB boot.

travis_san said:

What can I do to slipstream IE7 without generating missing file errors in text setup?

Try file names used by a default XP CD: that's 8.3 names.
Long names are renamed at installation.

txtsetup.sif said:

aaclient.mui = 100,,,,,,,,3,3,aaclient.dll.mui

The CD does contain file name aaclient.mui, not aaclient.dll.mui.

nLlite breaks XP defaults

txtsetup.sif said:

icardie.dll.mui = 1,,,,,,,1974,0,0

A long name is used at CD. Try a 8.3 file name at ISO image.

travis_san said:

I've been doing some testing with XP install files on the HD and it seems the errors I'm getting in Vmware are caused by an incorrect "SetupSourceDevice = " line in TXTSETUP.sif.

A XP default CD dosn't contain SetupSourceDevice, delete this line.
And if you use one hard disk, dosn't use grub4dos hd swap.

[doveman]

Actually neither of the ISOs I'm using contain the SetupSourceDevice line either. I just noticed that when I had the wrong settings in that line, whilst trying to do an install from the XPCD files copied to HD, that I got the same errors that I'm seeing with the ISO install.

I'm not using the HD swap when installing from the same HD either.

I had another problem in that the XP install was assigning C: to the third partion, but that turned out to just be because that partition was active. Setting the first partition active fixed that.

This post has been edited by doveman: 11 November 2009 - 10:58 AM

[travis_san]

Quote

NTLDR expects NTDETECT.COM in the active partition of the first disk, not on the second, as would be if no swapping is made, hence the need of swapping.

Thanks ilko_t, this works now.

Quote

There are several reports that nLited sources do NOT work well when booted not from a CD

To clarify, the nlited source installed via RAM load works fine. However, nlited source + IE7 slipstreamed creates errors in text mode.
I was considering cdob's post on how nlite breaks the naming convention in txtsetup.sif and yet this is functional if the source is an ISO created by nlite. I noticed these ISOs are Joliet, whereas XP_RAM.ISO was not. I simply added

set mkisofs_option=-R -J

to mkISO_RAMload_sort.cmd to enable Joliet. This works; no missing file errors in text mode.

Thank you for your help guys. I will test this on real hardware soon.

[doveman]

After reading travis_san's message, I thought "maybe I shouldn't be building my ISOs with Nero" (although it seems to make ISOs that work in the VCD-ROM) and recalled what I'd learnt about mkisofs only last week (and mostly forgotten already ).

Anyway, I rebuilt my "Vanilla" XP ISO with the below command, but it didn't help

mkisofs -v -volid XPPSP3 -R -l -iso-level 4 -J -joliet-long -duplicates-once -b boot.ima -no-emul-boot -boot-load-size 4 -boot-load-seg 0x07C0 -o XPPROSP3Un.iso "XPProSP3-Vanilla-Unattend"

[doveman]

To try and troubleshoot, I thought I'd just make an ISO, containing only XP_RAM.iso, XP_INST.IMA, grldr and menu.lst, to boot in VMware's VCD-ROM. It seems to be working OK, but I can't get it to respect my unattended settings in winnt.sif.

I've tried putting it in the root and in setup\I386\ of the IMA, but I still get the EULA every time. Where exactly am I meant to put this file?

[kDn]

doveman, on Nov 12 2009, 03:10 AM, said:

To try and troubleshoot, I thought I'd just make an ISO, containing only XP_RAM.iso, XP_INST.IMA, grldr and menu.lst, to boot in VMware's VCD-ROM. It seems to be working OK, but I can't get it to respect my unattended settings in winnt.sif.

I've tried putting it in the root and in setup\I386\ of the IMA, but I still get the EULA every time. Where exactly am I meant to put this file?

Try to reread Post #73

This post has been edited by kDn: 12 November 2009 - 07:01 AM

[doveman]

Thanks, but I'm not trying to do a multiple unattended-mode installation, so all that tells me is that it goes in the root of the IMA, which I've already tried.

I tested by copying the XP_RAM.iso and XP_INST.IMA to a usb stick and booting from that, starting the RAM install with drive mapping, and that worked fine.

It turns out that it's just a problem with my test CD ISO, to do with case. I don't even have to put WINNT.SIF in the IMA, it works fine from the XP_RAM.ISO once the problem's fixed.

After a lot of testing, it turns out that XP_RAM.iso needs to be XP_RAM.ISO for the WINNT.SIF to work from a CD ISO in Vmware. I've got no idea why it should work with a lower-case extension from a USB stick but not a CD ISO but there you are.

[cdob]

travis_san, on Nov 11 2009, 04:01 PM, said:

this is functional if the source is an ISO created by nlite. I noticed these ISOs are Joliet, whereas XP_RAM.ISO was not.

Sorry I didn't understand there are different ISO files. I understood you used the nLite ISO image always.

To allow multiple dots at ISO 9660: mkisofs -allow-multidot
Yes, or add Joliet file system: mkISO_RAMload_sort.cmd -joliet-long
Rock Ridge extension won't matter in that case. Anyay use -r instead of -R at windows in future.

doveman, on Nov 12 2009, 09:43 AM, said:

It turns out that it's just a problem with my test CD ISO, to do with case. I don't even have to put WINNT.SIF in the IMA, it works fine from the XP_RAM.ISO once the problem's fixed.

Yes, winnt.sif is not WINNT.SIF.

Quote

After a lot of testing, it turns out that XP_RAM.iso needs to be XP_RAM.ISO for the WINNT.SIF to work from a CD ISO in Vmware. I've got no idea why it should work with a lower-case extension from a USB stick but not a CD ISO but there you are.

Grub4dos ignore case at (floppy, hard) disk file system, your USB stick.
Grub4dos is case sensitive at ISO 9660 file system, your CD.

[doveman]

Quote

Yes, winnt.sif is not WINNT.SIF.

I didn't have any problems with that, the problem was purely with XP_RAM.iso .

Quote

Grub4dos ignore case at (floppy, hard) disk file system, your USB stick.
Grub4dos is case sensitive at ISO 9660 file system, your CD.

Yeah, it turns out that Grub4dos was actually finding the XP_RAM.iso on my VHD third partition, so was set root and loading XP_RAM.iso from there, which was an old, incorrectly built one, which wasn't loading WINNT.SIF. I've got round that problem by changing the first line to "find --set-root /XPRAMCD.lst" and making that file in the root of the CD. Very confusing, but I've finally managed to successfully install from USB, CD and HD, in VMware at least

[doveman]

Just wanted to report that thanks to TigerWhite's post #26, I've successfully installed from an Nlited Driverpacked XP_RAM.ISO contained in an ISO mounted in VMware's VCD-ROM. I was getting an error about unable to find firadisk in TXT-mode, until I added the line to mount the IMA to (fd1).

Only downside is it isn't installing the stuff in the I386\SVPACK folder for some reason, but pretty good job nonetheless.

[kof94]

A few years ago I remember someone saying that this sort of thing would never be possible with XP, so when I saw this I had to have ago.

Everything seams pretty straight forward and I'll test installing XP from a USB stick in the next day or so but I have one question.

Is there any additional steps required to mount a PE (UBCD4Win) in RAM using this method?

[doveman]

Someone may well correct me, but I'm not sure you need to go through all this just to boot a PE ISO in RAM.

I've had success using the following menu.lst entries. I think you just need to make sure your PE build includes the BootSDI plugin.

title PE iso
find --set-root /iso/PE.iso
map /iso/PE.iso (hd32)
map --hook
chainloader (hd32)

title PE.iso - mem
find --set-root /iso/PE.iso
map --mem /iso/PE.iso (hd32)
map --hook
chainloader (hd32)

[cdob]

kof94, on Nov 16 2009, 05:01 AM, said:

Is there any additional steps required to mount a PE (UBCD4Win) in RAM using this method?

No, there are no addional steps required.
However it's reccomended to "gzip.exe -9 UBCD4WIN.ISO" Required disk space and boot time is reduced.
http://www.gzip.org/ and ubcd4win\plugin\Add-Ons\UnxUtils\files\gzip.exe

title mem Loading UBCD4Win.GZ /I386/SETUPLDR.BIN
ls /images/UBCD4WIN.GZ || find --set-root /images/UBCD4WIN.GZ
map --mem /images/firadisk.ima (fd0)
map --mem /images/UBCD4WIN.GZ (0xFF)
map --hook
chainloader (0xFF)/I386/SETUPLDR.BIN

doveman, on Nov 16 2009, 05:43 AM, said:

Someone may well correct me, but I'm not sure you need to go through all this just to boot a PE ISO in RAM.

A classic XP/2003 won't boot without additional drivers.
Yes, a BootSDI version does work, because this use RAM load already (2003 SP1 setupldr.bin based).

[kof94]

Thank you.

This post has been edited by kof94: 16 November 2009 - 04:10 PM

[kof94]

Ok, I'm having the same issue as Doveman. After booting from hdd (and after presetup.cmd runs) I get an error that the installation CD cannot be found.

At best guess I'd say it hasn't been mounted at all.

I've trieed this in Qemu and on real hardware, same result... any ideas?

[doveman]

Most of my problems were because of the wrong case in filenames (XP_RAM.ISO, WINNT.SIF, DOSNET.SIF, TXTSETUP.SIF) so I'd check they're all upper-case as a first-step.

[jaclaz]

doveman, on Nov 16 2009, 11:43 AM, said:

Someone may well correct me, but I'm not sure you need to go through all this just to boot a PE ISO in RAM.

I've had success using the following menu.lst entries. I think you just need to make sure your PE build includes the BootSDI plugin.

I will gladly.

Booting a PE in RAM is not anything new.

It is possible since the release of Server2003 SP1, end of 2004/first months of 2005 methods were found:
http://www.911cd.net...?showtopic=9685
http://www.911cd.net...showtopic=10445
http://www.911cd.net...showtopic=10482
http://www.911cd.net...showtopic=10806

The BootSDI plugin is just the way some winbuilder projects implemented RAMDISK booting.

The way you depict, (mapping the iso to --mem with grub4dos) you are effectively loading the PE in memory TWICE, read here WHY:
http://www.boot-land.net/forums/index.php?...=8258&st=37

With Firadisk or WinVblock, you re-use the SAME image in RAM loaded via grub4dos.

jaclaz

[kof94]

My bad, I gzipped the XP iso's and ImDisk obviously can't handle that!