asqasq commented on a change in pull request #72: Allow namenode as parameter
instead of conf file.
URL: https://github.com/apache/incubator-crail/pull/72#discussion_r265112474
##########
File path: client/src/main/java/org/apache/crail/conf/CrailConfiguration.java
##########
@@ -39,6 +39,17 @@ public CrailConfiguration() throws IOException {
mergeProperties(properties);
}
+ // This constructor is useful, when a client does
+ // not have access to a file system, but only needs the configuration
+ // of the namenode address. As an example, an OpenWHisk function
+ // does not have a filesystem associated and therefore cannot
+ // read the namenode address from any configuration file.
+ public CrailConfiguration(String nameNodeAddress) throws IOException {
+ conf = new ConcurrentHashMap<>();
+ conf.put("crail.namenode.address", nameNodeAddress);
Review comment:
Good point, actually I was thinking about that as well... If we change it
this way, the default empty constructor would not load the properties, which
will break existing code. This is why I originally introduced a different
(simplistic) constructor. With this one we would not have the problem of
breaking or needing to change existing code.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
