I want to confirm if my understanding with Hadoop FileSystem object is correct or not. From the source code of org.apache.hadoop.fs. FileSystem (either from version 1.0.4 or 2.2.0), the method
public static FileSystem get(URI uri, Configuration conf) throws IOException
is using some sort of cache:
CACHE.get(uri, conf);
My understanding is that Tomcat usually create multiple threads to handle Http requests, and those threads will use the same FileSystem object (because of the cache).
This will resulting in an error, right?
The next question is, if I want to disable the cache, should I just introduce a new key "fs.hdfs.impl.disable.cache" and set the value to "true"?
And another key "fs.har.impl.disable.cache" for HAR FileSystem?