[
https://issues.apache.org/jira/browse/BROOKLYN-548?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Thomas Bouron updated BROOKLYN-548:
-----------------------------------
Description:
I was testing Brooklyn HA and noticed something odd: when restarting an HA
Brooklyn node, it changes it's node ID reported by `/v1/server/ha/states`.
For instance, I have 2 Brooklyn nodes, 1 master, 1 slave. I stopped master (not
killed, stopped), slave took over as expected. Then I restarted the previous
master, it came back online as a slave (expected) but the node ID was
different. I was expecting it to keep its ID rather creating a new one.
{code}
{
"planeId": "qDpn2Lau",
"ownId": "nJagUrw8",
"masterId": "nJagUrw8",
"nodes": {
"BgOjYcml": {
"nodeId": "BgOjYcml",
"nodeUri": null,
"status": "TERMINATED",
"localTimestamp": 1510052259169,
"remoteTimestamp": 1510052260000
},
"iv3XfaEu": {
"nodeId": "iv3XfaEu",
"nodeUri": null,
"status": "STANDBY",
"localTimestamp": 1510054973636,
"remoteTimestamp": 1510054974000
},
"nJagUrw8": {
"nodeId": "nJagUrw8",
"nodeUri": null,
"status": "MASTER",
"localTimestamp": 1510054975025,
"remoteTimestamp": 1510054976000
}
},
"links": {}
}
{code}
On the JSON above, {{BgOjYcml}} is the same node as {{iv3XfaEu}} but
{{BgOjYcml}} was the ID of the node before restart.
was:
I was testing Brooklyn HA and noticed something odd: when restarting an HA
Brooklyn node, it changes it's node ID reported by `/v1/server/ha/states`.
For instance, I have 2 Brooklyn nodes, 1 master, 1 slave. I stopped master (not
killed, stopped), slave took over as expected. Then I restarted the previous
master, it came back online as a slave (expected) but the node ID was
different. I was expecting it to keep its ID rather creating a new one.
{code}
{
"planeId": "qDpn2Lau",
"ownId": "nJagUrw8",
"masterId": "nJagUrw8",
"nodes": {
"BgOjYcml": {
"nodeId": "BgOjYcml",
"nodeUri": null,
"status": "TERMINATED",
"localTimestamp": 1510052259169,
"remoteTimestamp": 1510052260000
},
"iv3XfaEu": {
"nodeId": "iv3XfaEu",
"nodeUri": null,
"status": "STANDBY",
"localTimestamp": 1510054973636,
"remoteTimestamp": 1510054974000
},
"nJagUrw8": {
"nodeId": "nJagUrw8",
"nodeUri": null,
"status": "MASTER",
"localTimestamp": 1510054975025,
"remoteTimestamp": 1510054976000
}
},
"links": {}
}
{code}
On the attached screenshot, {{BgOjYcml}} is the same node as {{iv3XfaEu}}.
{{BgOjYcml}} was the ID of the node before restart
> Brooklyn node ID changes on restart
> -----------------------------------
>
> Key: BROOKLYN-548
> URL: https://issues.apache.org/jira/browse/BROOKLYN-548
> Project: Brooklyn
> Issue Type: Bug
> Affects Versions: 0.12.0
> Reporter: Thomas Bouron
>
> I was testing Brooklyn HA and noticed something odd: when restarting an HA
> Brooklyn node, it changes it's node ID reported by `/v1/server/ha/states`.
> For instance, I have 2 Brooklyn nodes, 1 master, 1 slave. I stopped master
> (not killed, stopped), slave took over as expected. Then I restarted the
> previous master, it came back online as a slave (expected) but the node ID
> was different. I was expecting it to keep its ID rather creating a new one.
> {code}
> {
> "planeId": "qDpn2Lau",
> "ownId": "nJagUrw8",
> "masterId": "nJagUrw8",
> "nodes": {
> "BgOjYcml": {
> "nodeId": "BgOjYcml",
> "nodeUri": null,
> "status": "TERMINATED",
> "localTimestamp": 1510052259169,
> "remoteTimestamp": 1510052260000
> },
> "iv3XfaEu": {
> "nodeId": "iv3XfaEu",
> "nodeUri": null,
> "status": "STANDBY",
> "localTimestamp": 1510054973636,
> "remoteTimestamp": 1510054974000
> },
> "nJagUrw8": {
> "nodeId": "nJagUrw8",
> "nodeUri": null,
> "status": "MASTER",
> "localTimestamp": 1510054975025,
> "remoteTimestamp": 1510054976000
> }
> },
> "links": {}
> }
> {code}
> On the JSON above, {{BgOjYcml}} is the same node as {{iv3XfaEu}} but
> {{BgOjYcml}} was the ID of the node before restart.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
