as johnpad mentioned, the update may fail and lead to an endless reboot if rotation is installed. if so, does uninstalling the rotation ensure a "seamless" update? i think many users like me just want to be reassured about this before going for the update..
I'm not aware of any evidence to support that. AFAIK, all existing rotation packages prevent the update from starting in the first place. The main culprit for reboot loops seems to be the preinstalled documentation and media.
When you have a non-booting system with a HDD, you can pull it, put it in another system, mount it read-only, and go look at logfiles, etc.
With other systems (e.g. Debian), this isn't needed: boot messages are output to the screen, so that one can see what's going on and possible errors. This is not always convenient, but still much better than having nothing at all like on the tablet. Moreover, one can enter single-user mode from the firmware, so that one can try to fix things.
There's also some tools that may be installed in the initfs to give you a USB networking (with telnet/ssh) or USB serial connection. This gets you a shell before it loops and you can diagnose/repair things from there.
Sounds really interesting. Any more information, please?
While there are some issues with this update, they mainly boil down to packaging and package-deployment issues. There is no real inability to diagnose failures, and anyone who can't clone to SD probably couldn't diagnose the failure in any case.
No, this needs to anticipate. One should really be able to diagnose the failure without anticipating and without having a spare SD. This isn't much a problem for this failure as it occurs for many users, so that some of them can diagnose it. But some other uncommon failures could really be hard to diagnose.
Well,... I for one feel the need to defend this guy... Identify the specific causes?!?!?! How on Earth is anyone supposed to do that when the darn thing keeps rebooting in an endless cycle???
Something is different on the tablet of every user for whom it's failed. No-one's going to be able to pin down what's common if what's been done to them isn't clear.
"Specific causes" meant - if it's failed for you - what, exactly have you done to the tablet since you last reflashed it? Saying "nothing, it's the same" and then later it turns out "well, I installed sliderotate" (not mentioning any particular user, here) means something was done.
Software doesn't fail randomly, unless there's a hardware failure. Something is different which is causing reboot loops etc.
Then, because I'd deleted stuff out of MyDocs, and put links there to the internal mmc card, the document package didn't work, I had to remove the links and re-create the dirs and put some dummy files in. Finally, the packages installed and dpkg-reconfigure worked, and I was upgraded.
I had the same problem. I had .documents linked to a folder on the internal mmc card, and the update process kept failing. The log read: mv: cannot rename '/usr/share/pre-installed/MyDocs/.documents/*': No such file or directory
After creating a new folder .documents under /usr/share/pre-installed/MyDocs with some dummy file in it, I could finally install the update successfully.
Some users said that the SSU broke their bootmenu, so how would this solve anything concerning the endless reboot loop?
Moreover this wiki page doesn't mention the tools to get USB networking before the reboot (http://wiki.maemo.org/USB_networking is silent on this point too).
Software doesn't fail randomly, unless there's a hardware failure. Something is different which is causing reboot loops etc.
Unfortunately not everyone remembers everything what they did since the last reflash, including things not done as root (as the SSU behavior also depends on user space, as it was shown).
I cannot understand why SSU installs anything to user's home directory at all. Home directory is for user to do anything he wants and trying to make a deb which works there 100% is not a easy task.
