I have a fairly busy site hosted with a rackspace cloud site.  It is currently running 5.1.4.
I copied the site to a local server to test the upgrade to 5.5.1, the upgrade went flawlessly.
I think backedup my entire site and database (thankfully), set app_offline.htm and uploaded the upgrade (is there a better way on this?)
I have autoupgrade set to false.
I then removed the app_offline.htm and accessed the upgrade url.
It ran for a bit and got about halfway through the upgrade process and timed out.
I hit the site again, got a bit further, this went on 3 or 4 times and eventually got to 'upgrade completed'
When I got to access your site link, I was never able to access the main page, I simply got asp error pages.
I looked in the event log (directly via SQL tools), the only thing that I noted was errors about POET vulnerability.
At this point, I was under the gun, the site had been down for 2 hours and I needed to restore from backup.

1) has anyone who hosts their site at rackspace (or any other shared host) successfully upgraded
2) I suspect I'm hitting timouts?  is there a way to make them longer?
3) is there a better way to upgrade when you're moving files around w/FTP?  (yes, it's a PIA)
4) should I simply pull the site local, upgrade it and then push it holesale back up?  (seems a bit extreme to me)

Thank you very much for your advice

That really isn't the answer that I'm looking for. As doing that means that either I have to take the site down for 2 or 3 hours (file transfers back and forth) or accept dataloss. Neither is particularly appetizing.

I'd really really like to a proper inplace upgrade as it will be the least disruptive.

Any thoughts on trying smaller steps and doing the upgrade as a series of smaller chunks? Or am I asking for trouble if I do that?

It seems that the upgrade ALWAYS craps out after 30 seconds or so. Is it possible to simply make the timeout longer? 30 seconds seems unreasonably short. It's nice on a developer's workstation where things happen quickly, but in a shared hosting environment, extensive operations like an upgrade can take a little time.

I reran on a test instance @ rackspace. This is my output:
Upgrade Status Report 00:00:00.390 - Upgrading to Version: 05.06.00
00:00:00.608 - Executing Script: 05.02.00.SqlDataProvider Success
00:00:05.881 - Executing Script: 05.02.01.SqlDataProvider Success
00:00:06.442 - Executing Script: 05.02.02.SqlDataProvider Success
00:00:06.661 - Executing Script: 05.02.03.SqlDataProvider Success
00:00:06.926 - Executing Script: 05.03.00.SqlDataProvider Success
00:00:10.374 - Executing Script: 05.03.01.SqlDataProvider Success
00:00:10.514 - Executing Script: 05.04.00.SqlDataProvider Success
00:00:12.215 - Executing Script: 05.04.01.SqlDataProvider Success
00:00:12.449 - Executing Script: 05.04.02.SqlDataProvider Success
00:00:12.776 - Executing Script: 05.04.03.SqlDataProvider Success
00:00:12.979 - Executing Script: 05.04.04.SqlDataProvider Success
00:00:13.104 - Executing Script: 05.05.00.SqlDataProvider Success
00:00:17.846 - Executing Script: 05.05.01.SqlDataProvider Success
00:00:18.033 - Executing Script: 05.06.00.SqlDataProvider Success
00:00:20.592 - Executing Application Upgrades: 05.02.00 Success
00:00:34.710 - Executing Application Upgrades: 05.02.01 Success
00:00:34.788 - Executing Application Upgrades: 05.02.02 Success
00:00:34.788 - Executing Application Upgrades: 05.02.03 Success
00:00:34.788 - Executing Application Upgrades: 05.0

 yes, it stops mid line. the last entry in the event log is a HOST_ALERT, there are NO ERRORS.
This is the text of the last entry in the event log:   UpgradeConfig Remove System Web Extensions

 I've grabbed an image of the file system at this stage, is there a log that I can look in to see what happened?