destroy-machine --force
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| | juju-core |
High
|
William Reade | ||
| | 1.16 |
Critical
|
William Reade | ||
| | juju-core (Ubuntu) |
Undecided
|
Unassigned | ||
| | Saucy |
Undecided
|
Unassigned | ||
Bug Description
[Impact]
Unresponsive machines cannot be removed/destroyed, consuming resourcs.
[Test Case]
This of course requires a unresponsive machine; however this can be simulated by destroying a machine that is still installing.
juju destroy-machine <machineid> # will fail
juju destroy-machine --force <machineid> # will succeed
[Regression potential]
Part of the upstream tested 1.16.6 point release.
[Original bug report]
Unresponsive machine agents may require that we forcibly set the machine to Dead to allow the provisioning agent to release its instance.
see also lp:1089289
Related branches
- juju hackers: Pending requested 2013-11-11
-
Diff: 1119 lines (+369/-287)17 files modifiedcmd/juju/destroymachine.go (+18/-1)
cmd/juju/destroymachine_test.go (+51/-14)
cmd/juju/destroyunit.go (+1/-1)
cmd/jujud/machine_test.go (+1/-1)
state/api/client.go (+7/-1)
state/api/params/params.go (+1/-0)
state/apiserver/client/client.go (+52/-2)
state/apiserver/client/client_test.go (+136/-31)
state/apiserver/client/perm_test.go (+1/-1)
state/cleanup_test.go (+26/-8)
state/life.go (+8/-1)
state/life_test.go (+5/-15)
state/machine.go (+20/-3)
state/machine_test.go (+13/-44)
state/service_test.go (+3/-3)
state/state.go (+0/-67)
state/unit_test.go (+26/-94)
| William Reade (fwereade) wrote : | #1 |
| Changed in juju-core: | |
| status: | New → Triaged |
| importance: | Undecided → Low |
| William Reade (fwereade) wrote : | #2 |
...but still, actually, high priority; just that it should be done after the unit is, because a replacement for a zombie machine generally cannot act usefully without replacements for its assigned zombie units in play.
See also lp:1190715
| Changed in juju-core: | |
| importance: | Low → High |
| Changed in juju-core: | |
| milestone: | none → 1.16.0 |
| summary: |
- terminate-machine --force + destroy-machine --force |
| Mark Ramm (mark-ramm) wrote : | #3 |
Is this related to the bug kapil filed?
| Changed in juju-core: | |
| milestone: | 1.16.0 → 2.0 |
| tags: | added: canonical-webops |
| tags: | added: destroy-machine |
| tags: | added: theme-oil |
| Curtis Hovey (sinzui) wrote : | #4 |
See bug 1233457 about a cases where the state-server failed to handle a cases where a dying agent's machine terminated before the state server cold tell it that it was dead.
| Changed in juju-core: | |
| importance: | High → Critical |
| milestone: | 2.0 → 1.17.0 |
| Changed in juju-core: | |
| assignee: | nobody → William Reade (fwereade) |
| status: | Triaged → In Progress |
| Changed in juju-core: | |
| milestone: | 1.17.0 → 2.0 |
| Changed in juju-core: | |
| milestone: | 2.0 → 1.17.0 |
| Changed in juju-core: | |
| status: | In Progress → Fix Committed |
| John A Meinel (jameinel) wrote : | #5 |
This is something we'd like to consider backporting to the 1.16 series. We may decide not to, but we don't want to forget about it.
| Kapil Thangavelu (hazmat) wrote : | #6 |
so for folks on 1.16.3 i've had put together a script that does the direct db deletion. it should only be considered after db introspection as its specifically around the scenario of a machine with no units, stuck in dying. it comes with no warranty, if you need to use it please ping me on irc before hand.
| Kapil Thangavelu (hazmat) wrote : | #7 |
Script to dump db for for foresenic analysis, used to verify db structure before use of machine removal script.
| John A Meinel (jameinel) wrote : | #8 |
The changes for this got bumped off to the side in my lp:~jameinel/juju-core/preparation-for-1.16.5
So I'm retargetting this for 1.16.5.
| John A Meinel (jameinel) wrote : | #9 |
The changes were landed back into the 1.16 series.
| Changed in juju-core: | |
| importance: | Critical → High |
| Changed in juju-core: | |
| status: | Fix Committed → Fix Released |
| Ubuntu QA Website (ubuntuqa) wrote : | #10 |
This bug has been reported on the Ubuntu ISO testing tracker.
A list of all reports related to this bug can be found here:
http://
| tags: | added: iso-testing |
| Changed in juju-core (Ubuntu): | |
| status: | New → Fix Released |
| description: | updated |
| Rolf Leggewie (r0lf) wrote : | #11 |
saucy has seen the end of its life and is no longer receiving any updates. Marking the saucy task for this ticket as "Won't Fix".
| Changed in juju-core (Ubuntu Saucy): | |
| status: | New → Won't Fix |
I consider this low priority until we have the unit story sorted out cleanly.