Merge lp:~allenap/maas/no-such-cluster-for-all-the-things into lp:~maas-committers/maas/trunk
Proposed by
Raphaël Badin
Status: | Merged | ||||
---|---|---|---|---|---|
Approved by: | Gavin Panella | ||||
Approved revision: | no longer in the source branch. | ||||
Merged at revision: | 3163 | ||||
Proposed branch: | lp:~allenap/maas/no-such-cluster-for-all-the-things | ||||
Merge into: | lp:~maas-committers/maas/trunk | ||||
Diff against target: |
74 lines (+23/-7) 3 files modified
src/maasserver/rpc/leases.py (+13/-6) src/maasserver/rpc/tests/test_regionservice.py (+7/-0) src/provisioningserver/rpc/region.py (+3/-1) |
||||
To merge this branch: | bzr merge lp:~allenap/maas/no-such-cluster-for-all-the-things | ||||
Related bugs: |
|
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Newell Jensen (community) | Approve | ||
Review via email: mp+236797@code.launchpad.net |
Commit message
When update_leases() does not find the given cluster, raise NoSuchCluster.
Previously it was raising a Django HTTP 404 error.
To post a comment you must log in.
I was seeing logs like the following in maas.log when starting from a
clean-slate (i.e. no auto-accepted master cluster yet exists in the
database):
maas. lease_upload_ service: [ERROR] Failed to upload leases:
Code<UNKNOWN>: Unknown Error
The RPC connection was dropped after this because that's what AMP does:
it drops connections when unrecognised exceptions are transmitted. The
cluster would eventually be registered, but there can be more failures
before it does.
With this branch I get:
maas. lease_upload_ service: [ERROR] Failed to upload leases: The 244e-4144- 8cfc-b4dbc8cba1 88 could not be found.
cluster (a.k.a. node-group) with UUID
d791c8c0-
and the RPC connection does not drop. This means that when controller (upstart job: maas-cluster- register) wakes up
start-cluster-
(every 60 seconds) there's a good chance that the cluster will be
connected to the region, and the registration will be successful.