2.2.1 runs 2 leadership workers
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Won't Fix
|
Low
|
Unassigned | ||
2.2 |
Won't Fix
|
Low
|
Unassigned |
Bug Description
I was testing out other issues wrt transactions and having a broken transaction in the queue.
I explicitly injected an invalid transaction token in order to test how mgopurge handles large transaction queues.
$ juju bootstrap lxd
$ juju deploy -m controller ubuntu --to 0
$ lxc exec juju-xxx bash
$$ dialmgo
> db.leases.
> db.leases.
I was then watching 'juju debug-log' and saw this:
machine-0: 17:48:55 INFO juju.worker start "leadership"
machine-0: 17:48:55 INFO juju.worker start "leadership"
machine-0: 17:48:55 INFO juju.worker stopped "leadership", err: cannot find transaction ObjectIdHex(
machine-0: 17:48:55 ERROR juju.worker exited "leadership": cannot find transaction ObjectIdHex(
machine-0: 17:48:55 INFO juju.worker restarting "leadership" in 1s
machine-0: 17:48:55 INFO juju.worker stopped "leadership", err: cannot find transaction ObjectIdHex(
machine-0: 17:48:55 ERROR juju.worker exited "leadership": cannot find transaction ObjectIdHex(
machine-0: 17:48:55 INFO juju.worker restarting "leadership" in 1s
machine-0: 17:48:55 ERROR juju.worker.
machine-0: 17:48:56 INFO juju.worker start "leadership"
machine-0: 17:48:56 INFO juju.worker start "leadership"
machine-0: 17:48:56 INFO juju.worker stopped "leadership", err: cannot find transaction ObjectIdHex(
machine-0: 17:48:56 ERROR juju.worker exited "leadership": cannot find transaction ObjectIdHex(
machine-0: 17:48:56 INFO juju.worker restarting "leadership" in 1s
machine-0: 17:48:56 INFO juju.worker stopped "leadership", err: cannot find transaction ObjectIdHex(
machine-0: 17:48:56 ERROR juju.worker exited "leadership": cannot find transaction ObjectIdHex(
machine-0: 17:48:56 INFO juju.worker restarting "leadership" in 1s
machine-0: 17:48:57 INFO juju.worker start "leadership"
machine-0: 17:48:57 INFO juju.worker start "leadership"
unit-ul-0: 17:48:57 ERROR juju.worker.
unit-ul2-2: 17:48:57 ERROR juju.worker.
machine-0: 17:48:57 INFO juju.worker stopped "leadership", err: cannot find transaction ObjectIdHex(
machine-0: 17:48:57 ERROR juju.worker exited "leadership": cannot find transaction ObjectIdHex(
machine-0: 17:48:57 INFO juju.worker restarting "leadership" in 1s
machine-0: 17:48:57 INFO juju.worker stopped "leadership", err: cannot find transaction ObjectIdHex(
machine-0: 17:48:57 ERROR juju.worker exited "leadership": cannot find transaction ObjectIdHex(
machine-0: 17:48:57 INFO juju.worker restarting "leadership" in 1s
To start with, I thought we restarted at 3s not 1s, but even with that, you can see that there are 2 messages about restarting the 'leadership' worker within 1s. Such as:
machine-0: 17:48:57 INFO juju.worker start "leadership"
machine-0: 17:48:57 INFO juju.worker start "leadership"
This may only be because I'm messing with the 'controller' model. I have the feeling we might be running all the workers for the Controller *and* for the controller Model.
Yes, we run the leadership workers for each state. I believe the leadership worker is only interested in docs for its own model.