[
https://issues.apache.org/jira/browse/OMID-123?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16734431#comment-16734431
]
Lars Hofhansl commented on OMID-123:
------------------------------------
I see. Thanks [~ohads] that makes sense!
How do ZK and HBase do that? Tephra also has an HA mode. The ZK quorum peers
have to know about each other, perhaps we can borrow from there...?
The simple patch here would still provide for that, right? (Tries to find an
en/eth interface, if none found it uses the first inet interface).
> Allow starting Omid on machines without en or eth interfaces.
> -------------------------------------------------------------
>
> Key: OMID-123
> URL: https://issues.apache.org/jira/browse/OMID-123
> Project: Apache Omid
> Issue Type: Improvement
> Reporter: Lars Hofhansl
> Assignee: Yonatan Gottesman
> Priority: Major
> Attachments: OMID-123.txt
>
>
> Currently there's an assumption that there must be an eth* or a configured
> en* interface in order to bind to a specific address.
> {code}
> Exception in thread "main" java.lang.IllegalArgumentException: No network
> 'en*'/'eth*' interfaces found
> at
> org.apache.omid.NetworkUtils.getDefaultNetworkInterface(NetworkUtils.java:52)
> at org.apache.omid.tso.TSOServerConfig.<init>(TSOServerConfig.java:88)
> at org.apache.omid.tso.TSOServerConfig.<init>(TSOServerConfig.java:56)
> at org.apache.omid.tso.TSOServer.main(TSOServer.java:147)
> {code}
> The first question is: Why don't we just allow to bind to ::?
> Second, we can at least provide a default way to start when no en/eth
> interface is available.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
