8.10 ubiquity fails during partitioning

[Victor Churchill]

Binary package hint: ubiquity

I do not have a traceback for this.
This laptop has a Windows XP partition on it. I have asked for a 'guided install', resizing scsi1(0,0,0) and using freed space.
I think the problem may be related to the presence of a "internal to Windows" installation of Ubuntu 8.04 from several months ago. That could be consistent with the ntfsresize error reported in /var/log/messages (attached)

[hillbilly1980 (keith-green-light)]

This problem occurs with an HP ML350 Server while i try and install ubuntu 8.10 server after already having installed windows 2003 SBS server. There are two raids on the controller, a 146GB raid 1+0 and a 500approx raid 5. and a second raid control that is simply being used to provide access to a few 500gb sata drives.

When i try and do the guided install it fails b/c it only gives an option to use a second raid controller, i can't choose, regardless i run down to "manual" and hit okay.

I get the following listed:

SCSI.CCB (-,0,0) (cciss/c0d0) - 146.8 GB Compaq Smart Array
                   #1 primary 146.8 GB B ntfs
                        pri/log 8.4mb Free Space
SCSI.CCB(-,0,1) (cciss/c0d1) - 500 GB Compaq Smart Array
....

I wont' get into the second and 3rd disks, the second does the same thing as the first and follows the same listing as the first anyway, just a different size. The 3rd is just one big disk, don't know if it causes a problem, i don't want ubuntu on there so i'm not wasting my time with it :).

When i select the primary and then select "resize partition" i get a red error page advising me to visit the console 4. Console 4 fails giving me an error like.

/dev/cciss/c0d0 Invalid startup volume... ntfsresize --info failed disk does not have a valid NTFS..... I

Sorry i know i butched that... lazy yes...

If i run into the second console and go into /dev/cciss/c0d0 I see that the actual partition is labled /dev/cciss/c0d0p1. Not sure what problem it is, maybe partman but it's reading out the partion info correctly just not handing the entire partition path to ntfsresize. Which causes it to fail on this type of raid controller.

Let me know if you need more info.

[hillbilly1980 (keith-green-light)]

Just wanted to add, since this failed to do the ntfsresize i tried myself by setting my partion to 75G then going into fdisk and deleting the partion and recreating it as 75G. I didn't have much to lose if the windows blew up, it was more an excersize in doing it. I continued with the install by allocating my new free space to ubuntu back in the wizard in console 1 and finished the install, everything went fine until the reboot. Grub had found my windows parttion in the last step but then i was greated with an invalid disk.

I booted from a rescue disk and was able to mount the ext3 on part 4 and follow these instructions to grub running. Substituting my the device.map located on my ubuntu partion and menu.lst for the conf file. I ran the commands against part 4 and hd1 instead of hd0, considering my particular situation.

Root (hd1,4)
setup (hd1)

http://nozell.com/blog/2004/05/08/some-tips-to-reinstall-grub-on-an-hp-proliant-server/

My best guess is that grub must also have a problem with grabbing that last partition, just like ntfsresize, there were a number of quoted bugs regarding this online, a few mentioned it might be a recently introduced reversion

While my windows part won't boot, it is listed but says the ntldr is missing, i'm sure i can fix that later should i care to. Te files are intact when i mount that part in ubuntu, so i probably didn't do the resize incorrectly. Just in case your wondering.

[Phillip Susi (psusi)]

It looks like this was some missing support in ntfsresize. Given that this was over 5 years ago, and ntfs support has come a long way in that time, I'm pretty sure this has been fixed.