Handling Failure Gracefully

Registered by Robert Ancell on 2011-10-25

In the real world there are always going to be failures, triggered by
things like software bugs, hardware failures and misconfiguration.
Ubuntu should where possible handle common failures and provide
predictable feedback to the user that the system is broken.

Blueprint information

Status:
Not started
Approver:
Sebastien Bacher
Priority:
Medium
Drafter:
Robert Ancell
Direction:
Approved
Assignee:
Robert Ancell
Definition:
Approved
Series goal:
Accepted for quantal
Implementation:
Deferred
Milestone target:
milestone icon ubuntu-12.10-beta-1

Related branches

Sprints

Whiteboard

Work done in precise:
[robert-ancell] Produce list of things we can detect and actions to take (https://wiki.ubuntu.com/GracefulFailure): DONE
[bryce] get failsafe-x working: DONE

pitti, 2012-03-28: Moving blueprint to Q, as most stuff is still open.

mathieu-tl, 2012-07-04: ncurses network-manager not required; we already have nmcli which ships with NM and provides a way to create new connections for wireless networks now.

(?)

Work Items

Work items:
[cyphermox] investigate ncurses network manager connection: DONE

Work items for ubuntu-12.10-beta-2:
[ev] Ensure that for any of these failures we're recording apport crashes: POSTPONED
[ev] If a crash occurs and the user has crash database reporting turned on, automatically create the .upload file so that it gets uploaded the next time there is internet: POSTPONED
[ev] Work out methods for storing flags to detect boot failures: POSTPONED
[jamesodhunt] discuss with mvo+stgraber how we might pass data into friendly-recovery so we can display a message suggesting the user select a particular option: POSTPONED