Connecting to Ubuntu Linux using VNC only gives 640x480 resolution

xorg.conf has nothing to do with vnc! What has to do whict vnc is the Xvnc server (run on the Ubuntu side), which can (depending on parameters) limit the _virtual_display to a certain size, and the client (on some other host) which theoretically can manipulate what size viewport you have (few do though, most implementations aim at what UltraVNC does so nicely... scaling:-).

Unfortunately I've go limited experience with Ubuntu (I'm planning to start play with it the next few weeks... VMWare and PHB willing:-), but I imagine you need start looking at how Xvnc get spawned off (started)... Likely in /etc/init.d/vncserver or something similar... Usually the call is not directly to Xvnc, but rather to the vncserver conveniance script.

-- Glenn

>> you might start by editing /etc/vnc.conf and set "$geometry = 1024x768", then restart things and see what gives.
> I actually tried something like this, also on Breezy: i have my screen connected at 1024x768 and changed $geometry to "640x480 in vnc.conf... No change..
As expected, with vino;)

>> So for your theory to work hfern, gtomsho would need log in blind (without a screen attached)...
> I actually logged out and guess what: there is no way how you can log into the system over vnc using vino. You indeed need to log in first on thje console and only
> then will vino let yousetup a connection Again, this is different then vncserver which defines a virtual X-terminal for you to log into.
Yep, as stated:).

> I tried the blind login with my screen disconnected: on my system it still provides a 1024x768 screen over vino.
As expected... If the gfx adaper will work headless, it'll nog magically "downsize"... Unless it's a terminally stupid intel chipset gfx (I've not tested this, but those suckers poll the bios for available modes and might ... get confused...).

> This could be hardware related though, like Gns mentioned: 'Assuming the graphics adapter "wants to play" in such a setup'.
Perhaps, yes. But more likely SW, as a working hypothesis:-).

> I'd be interested to learn what gtomsho's testing with the monitor stilll in place will yield.
> Also, if gtomso reallly wants to leave the system without screen than this vino desktop sharing may not be the best solution. I would recommend a vncserver
> setup. This will give you all the flexibility to access the system remotely. Alternatively, connecting with an X-Terminal software package may be on option.

Yes, quite. Both on the recommendation to give us more info, and the recommendation to work with a real vnc (note: doesn't have to be RealVNC(tm):-) server setup... Which Ubuntu will gladly support.
You could even mix them, since the default setyup for any vncserver package is to provide a virtual screen at :1 (actually port 5901) while vino defaults to :0 (teh actual desktop.... on port 5900). "Best of both worlds":-).

As to determining what is up ... One has to check the .vnc directory (for any stray settings there) for the user Xvnc is running as, as well as the /etc/vnc.conf file...
Pretty much up to you now gtomsho, which way we go with this... Please work with us!

-- Glenn

Beg to differ. Although I wasn't up to speed at the begining of the Q, I certainly contribute to the possible solutions. This should be a split, if anything.

-- Glenn

Read my second comment and note that it precedes your second comment. I'm not saying you shouldn't have any credit, nor that you shouldn't be credited with the "answer" as such, only that you are not the sole contributor to the probable solution.
You didn't exactly nail this one either, but we did perhaps do so together.

-- Glenn

Exactly my point Andrew... As I see it, there is no "true solution" as of yet, and there certainly cannot be one without point 99.
Hfern, your comments are valuable, knowledgeable and probably useful. But do they constitute a true answer? Using "definite" language doesn't mean much in this (I tend to be ... cautious(sp?)... in my formulations... Doesn't mean I don't think I'm right, just that I'm not adamant:-):-). Your second comment corroborates that this is probably not vino at all (which concurs with my testing), but rather Xvnc some way or other... And then we're probably not "detailed enough", any one of us (in which case it might be right to just delete the Q).
My tests on a machine with intel chipset gfx has not revealed the behaviour, but that just shows that that particular machine doesn't exhibit this (with vino).

The most relevant (common to us all) comment is the plea for interraction from gtomsho.
And I'm certainly not out to "fight", nor to "hog a Q"... I'm just announcing a difference of opinion as to the disposition of the Q.

-- Glenn

Good suggestion. Who knows, it might still be HW.... Personally I hate these questions where one never reaches an "endpoint", regardless if I'm right or not.... (I've seen some things turn around on my reasoning, once the asker has taken the time to actually participate, more than once, to say the least:-).
We'll see, hopefully.

Cheers
-- Glenn