You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@whirr.apache.org by "Steve Loughran (JIRA)" <ji...@apache.org> on 2013/06/21 16:40:20 UTC

[jira] [Commented] (WHIRR-730) Use the InetAddress.getLocalHost() instead of URL("http://checkip.amazonaws.com/") to get the local address

    [ https://issues.apache.org/jira/browse/WHIRR-730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13690331#comment-13690331 ] 

Steve Loughran commented on WHIRR-730:
--------------------------------------

I don't know about " illegal for an application to get the client ip address" through an external web site -that's what AWS  provide it for. More relevant is is "of limited value in an Intranet, proxy-or-not, because you'll get the external firewall IPAddr, not anything internal.

However, before rushing to swap over to {{getLocalHost()}}, 
# On a multihome device you may get one of the hostnames seemingly at random. That may be precisely what you don't want on public cloud infrastructures.
# You need to add handlers for the situation where the machine doesn't know its own hostname or IP address. 

For issue #2, Hadoop's NetUtil package provides the (robust) methods needed

                
> Use the InetAddress.getLocalHost() instead of URL("http://checkip.amazonaws.com/") to get the local address
> -----------------------------------------------------------------------------------------------------------
>
>                 Key: WHIRR-730
>                 URL: https://issues.apache.org/jira/browse/WHIRR-730
>             Project: Whirr
>          Issue Type: Improvement
>          Components: core
>    Affects Versions: 0.8.2
>         Environment: Unbuntu, Openstack
>            Reporter: Jeremy Lv
>             Fix For: 0.9.0
>
>         Attachments: FirewallManager.java.patch
>
>
> Hi,all
> I have found the FirewallManager.getOriginatingIp() use the following method to get the ip address
> {code:title=FirewallManager.java|borderStyle=solid}
>   /**
>    * @return the IP address of the client on which this code is running.
>    * @throws IOException
>    */
>   private String getOriginatingIp() throws IOException {
>     if ("stub".equals(clusterSpec.getProvider())) {
>       return "62.217.232.123";
>     }
>     URL url = new URL("http://checkip.amazonaws.com/");
>     HttpURLConnection connection = (HttpURLConnection) url.openConnection();
>     connection.connect();
>     return IOUtils.toString(connection.getInputStream()).trim() + "/32";
>   }
> {code}
> I think we should use the InetAddress.getLocalHost() to get the ipaddress, here's some of my reason:
> 1). I think it is illegal for an application to get the client ip address throght the erternal website, what else if the amazon is broken and we can access the http://checkip.amazonaws.com/ or if the situation of the internet can't worked well, it will be failed to launch the cluster because we can't connect to the http://checkip.amazonaws.com/.
> 2). If any of the private company want to reuse the whirr project to create the cluster, Some of the company need to use the proxy to access the external website, Although the whirr project support to config the proxy, but it is still complicated for us to apply for the IT Department to authorize the proxy for the VM we have created by the whirr.
> 3). I think it is simple to use the InetAddress.getLocalHost() to get the local ip address
> Thanks
> -Jeremy Lv

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira