Can't destroy service -- juju status "failed" & unknown relation 0 error in log

Bug #1581537 reported by Casey Marshall
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Canonical Juju
Triaged
Medium
Unassigned

Bug Description

Juju 2.0, built from master at 0bd48e6.

Steps to reproduce:

I deployed an etcd charm with 3 units, related it to prometheus, everything worked, no problem.

Did `juju destroy-service etcd` and two of the etcd units died and went away. However, one of them is stuck and repeated `juju destroy-service etcd` commands do not remove it.

The unit agent log on the stuck unit shows:

2016-05-13 13:32:41 INFO juju.worker.uniter resolver.go:120 committing "relation-broken" hook
2016-05-13 13:32:42 INFO juju.worker.uniter uniter.go:355 unit "etcd/2" shutting down: committing operation "skip run relation-broken (0) hook": unknown relation: 0
2016-05-13 13:32:42 ERROR juju.worker.dependency engine.go:526 "uniter" manifold worker returned unexpected error: committing operation "skip run relation-broken (0) hook": unknown relation: 0
2016-05-13 13:32:42 INFO juju.worker.leadership tracker.go:182 etcd/2 will renew etcd leadership at 2016-05-13 13:33:11.869693571 +0000 UTC
2016-05-13 13:32:45 INFO worker.uniter.jujuc tools.go:20 ensure jujuc symlinks in /var/lib/juju/tools/unit-etcd-2
2016-05-13 13:32:45 INFO worker.uniter.jujuc tools.go:40 was a symlink, now looking at /var/lib/juju/tools/2.0-beta7.1-xenial-amd64
2016-05-13 13:32:45 INFO juju.worker.uniter.relation relations.go:431 joining relation "prometheus:scrape etcd:target"
2016-05-13 13:32:45 INFO juju.worker.uniter.relation relations.go:467 joined relation "prometheus:scrape etcd:target"
2016-05-13 13:32:45 INFO juju.worker.uniter uniter.go:164 unit "etcd/2" started
2016-05-13 13:32:45 INFO juju.worker.uniter uniter.go:203 hooks are retried true
2016-05-13 13:32:45 INFO juju.worker.uniter resolver.go:120 committing "relation-broken" hook
2016-05-13 13:32:45 INFO juju.worker.uniter uniter.go:355 unit "etcd/2" shutting down: committing operation "skip run relation-broken (0) hook": unknown relation: 0
2016-05-13 13:32:45 ERROR juju.worker.dependency engine.go:526 "uniter" manifold worker returned unexpected error: committing operation "skip run relation-broken (0) hook": unknown relation: 0
2016-05-13 13:32:48 INFO worker.uniter.jujuc tools.go:20 ensure jujuc symlinks in /var/lib/juju/tools/unit-etcd-2
2016-05-13 13:32:48 INFO worker.uniter.jujuc tools.go:40 was a symlink, now looking at /var/lib/juju/tools/2.0-beta7.1-xenial-amd64
2016-05-13 13:32:48 INFO juju.worker.uniter.relation relations.go:431 joining relation "prometheus:scrape etcd:target"
2016-05-13 13:32:48 INFO juju.worker.uniter.relation relations.go:467 joined relation "prometheus:scrape etcd:target"
2016-05-13 13:32:48 INFO juju.worker.uniter uniter.go:164 unit "etcd/2" started
2016-05-13 13:32:48 INFO juju.worker.uniter uniter.go:203 hooks are retried true
2016-05-13 13:32:48 INFO juju.worker.uniter resolver.go:120 committing "relation-broken" hook
2016-05-13 13:32:48 INFO juju.worker.uniter uniter.go:355 unit "etcd/2" shutting down: committing operation "skip run relation-broken (0) hook": unknown relation: 0
2016-05-13 13:32:48 ERROR juju.worker.dependency engine.go:526 "uniter" manifold worker returned unexpected error: committing operation "skip run relation-broken (0) hook": unknown relation: 0
2016-05-13 13:32:51 INFO worker.uniter.jujuc tools.go:20 ensure jujuc symlinks in /var/lib/juju/tools/unit-etcd-2
2016-05-13 13:32:51 INFO worker.uniter.jujuc tools.go:40 was a symlink, now looking at /var/lib/juju/tools/2.0-beta7.1-xenial-amd64
2016-05-13 13:32:51 INFO juju.worker.uniter.relation relations.go:431 joining relation "prometheus:scrape etcd:target"
2016-05-13 13:32:51 INFO juju.worker.uniter.relation relations.go:467 joined relation "prometheus:scrape etcd:target"
2016-05-13 13:32:51 INFO juju.worker.uniter uniter.go:164 unit "etcd/2" started
2016-05-13 13:32:51 INFO juju.worker.uniter uniter.go:203 hooks are retried true
2016-05-13 13:32:51 INFO juju.worker.uniter resolver.go:120 committing "relation-broken" hook
2016-05-13 13:32:51 INFO juju.worker.uniter uniter.go:355 unit "etcd/2" shutting down: committing operation "skip run relation-broken (0) hook": unknown relation: 0
2016-05-13 13:32:51 ERROR juju.worker.dependency engine.go:526 "uniter" manifold worker returned unexpected error: committing operation "skip run relation-broken (0) hook": unknown relation: 0

Revision history for this message
Casey Marshall (cmars) wrote :

I'm able to recover from this condition with `juju remove-machine --force`, but that `unknown relation: 0` error is probably indicative of some problems.

Revision history for this message
Cheryl Jennings (cherylj) wrote :

Casey - if you have logs from the controller, can you attach them?

Changed in juju-core:
status: New → Incomplete
Revision history for this message
Casey Marshall (cmars) wrote :

Here's the machine log from the controller: https://pastebin.canonical.com/156442/

Changed in juju-core:
status: Incomplete → New
Changed in juju-core:
importance: Undecided → Medium
status: New → Triaged
affects: juju-core → juju
Revision history for this message
Anastasia (anastasia-macmood) wrote :

This has been fixed in Juju 2.1x by the fix of bug # 1643430

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.