[
https://issues.apache.org/jira/browse/HDDS-1354?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17134263#comment-17134263
]
Marton Elek commented on HDDS-1354:
-----------------------------------
I think this is still a problem. In CM environments the hostname and the fully
qualified hostname is the same:
{code:java}
[nightly7x-1 /]# hostname
nightly7x-1.nightly7x.root.hwx.site
[nightly7x-1 /]# hostname -f
nightly7x-1.nightly7x.root.hwx.site {code}
But we couldn't assume that it will be true for all the environments.
> Kerberos principal configuration of OzoneManager doesn't use FQDN
> -----------------------------------------------------------------
>
> Key: HDDS-1354
> URL: https://issues.apache.org/jira/browse/HDDS-1354
> Project: Hadoop Distributed Data Store
> Issue Type: Bug
> Components: Security
> Affects Versions: 0.4.0
> Reporter: Marton Elek
> Assignee: Ajay Kumar
> Priority: Minor
> Labels: Triaged
>
> In the "*.kerberos.principal" settings hadoop supports the _HOST variable
> which is replaced to the fully qualified domain name.
> For example:
> {code}
> OZONE-SITE.XML_hdds.scm.kerberos.principal: "scm/[email protected]"
> {code}
> It works well with scm but for om it uses the hostname instead of the FQDN.
> (SCM uses the HddsServerUtil.getScmBlockClientBindAddress which uses the
> _bind_ address but the om uses the om rpc address).
> I would suggest to use the same behaviour for both SCM and OM.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
