Working with Services
Restarting services
In the unlikely case that a service gets stuck in an invalid state and does not restart automatically, you can do a manual service restart like this:
$ kubectl -n demo get pods
NAME READY STATUS RESTARTS AGE
engine-77f988f5bc-9dsbx 2/2 Running 0 3h17m
iam-5cf4c44df9-gt6vh 2/2 Running 0 3h17m
index-77886848dd-msj6q 1/1 Running 0 3h17m
influxdb-58df459db6-4xp72 1/1 Running 14 115d
inventory-6bf767c8d5-qccrm 2/2 Running 0 3h17m
mqtt-646bfc5cf7-lcpjd 1/1 Running 0 3h17m
nats-7f89678b77-jnnj8 1/1 Running 0 3h17m
registry-7b4b5cb585-tf7bk 2/2 Running 0 3h17m
th-api-5988c7cf4-95tpj 1/1 Running 0 3h17m
tsdb-d46964779-qsjqq 1/1 Running 0 3h17m
visualizer-7bfd79676d-hz6pc 1/1 Running 0 3h17m
web-ui-6d45775d4d-zrhrs 1/1 Running 0 3h17m
$ kubectl -n demo delete pod engine-77f988f5bc-9dsbx
pod "engine-77f988f5bc-9dsbx" deleted
$
This can also be done in a single command by selecting on a service's app
label using -l
:
Known issues
$ $ kubectl.exe delete pod -lapp=engine -n <tenant>
pod "engine-77f988f5bc-r4429" deleted
$
Uplinks are received, but not processed for decoding
This has been observed to happen rarely after cluster reorganization. Until a final solution has been found, it is recommended to restart the affected tenants' engine service:
$ $ kubectl.exe delete pod -lapp=engine -n <tenant>
pod "engine-xxxxxxxxxx-yyyyy" deleted
$
Data gets decoded, but not passed on via integrations
This has been observed to happen rarely after cluster reorganization. Until a final solution has been found, it is recommended to restart the affected tenants' engine service:
$ $ kubectl.exe delete pod -lapp=engine -n <tenant>
pod "engine-xxxxxxxxxx-yyyyy" deleted
$
Note that this is likely coming from the same root cause as the previous known issue.