[
https://issues.apache.org/jira/browse/HADOOP-1974?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vivek Ratan updated HADOOP-1974:
--------------------------------
Attachment: 1974_patch01
>From my comments in HADOOP_1970:
There are 2-3 methods that traverse the tree-like structure of Progress
methods, and at least two of them traverse (and obtain locks) in different
directions, hence the deadlock. (One of) the right solution is to obtain locks
in one direction only - so we lock when going downwards from the root node.
This happens in get(), getInternal(), and toString(). If you need to traverse
upwards towards the root (in complete()), you either release your lock before
getting your parent's (which is what I've chosen to do, since we don't need
transactional semantics), or you get locks in the same direction as other
traversal methods.
Another somewhat related issue is that Progress::get(), which is called quite
often, always traverses upwards to find the root of a structure. Since a node's
root never changes, it should be cached at each node. This certainly improves
performance for get(), but it also offers a synch mechanism should we ever need
to write code that needs to lock multiple nodes and traverse upwards towards
the root. In such a case, the methods can lock the root object to get sole
access control to the entire structure. We don't need this for now, but it's a
good mechanism to have for the future.
> Progress node should cache root object for faster progress computation
> ----------------------------------------------------------------------
>
> Key: HADOOP-1974
> URL: https://issues.apache.org/jira/browse/HADOOP-1974
> Project: Hadoop
> Issue Type: Improvement
> Reporter: Vivek Ratan
> Assignee: Vivek Ratan
> Priority: Minor
> Attachments: 1974_patch01
>
>
> In org.apache.hadoop.util.Progress.get(), we walk through the tree of objects
> to find the root of a node, before figuring out the progress. This approach
> is not optimized, especially since get() is called frequently. Each
> Progressnode should cache its root object, and this is easy to do since nodes
> do not change their parents.
> Keeping track of the root node is also useful in synchronization issues. [see
> HADOOP-1970 for more details]. The root node can be used to synchronize the
> entire structure for methods that need to traverse the tree in different
> directions and lock nodes along the way.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
