[
https://issues.apache.org/jira/browse/HBASE-21915?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16771996#comment-16771996
]
Josh Elser commented on HBASE-21915:
------------------------------------
{quote}Whats up w/ the call to tryOpen inside the unbuffer? Are we opening the
stream when we do this? If fresh open, an unbuffer makes sense? Thanks.
{quote}
{quote}Sounds like we are creating a new stream just to unbuffer, leaving any
possible previous existing stream open
{quote}
Yup! You are both totally right. This was me hacking something together to run
out the door ;). Let me polish this first patch, and put some thought into a
unit test (although, I'm not sure of one that wouldn't be contrived).
> FileLink$FileLinkInputStream doesn't implement CanUnbuffer
> ----------------------------------------------------------
>
> Key: HBASE-21915
> URL: https://issues.apache.org/jira/browse/HBASE-21915
> Project: HBase
> Issue Type: Bug
> Components: Filesystem Integration
> Reporter: Josh Elser
> Assignee: Josh Elser
> Priority: Major
> Attachments: HBASE-21915.001.patch
>
>
> FileLinkInputStream is an InputStream which handles the indirection of where
> the real HFile lives. This implementation is wrapped via
> FSDataInputStreamWrapper and is transparent when it's being used by a caller.
> Often, we have an FSDataInputStreamWrapper wrapping a FileLinkInputStream
> which wraps an FSDataInputStream.
> The problem is that FileLinkInputStream does not implement the
> \{{CanUnbuffer}} interface, which means that the underlying
> {{FSDataInputStream}} for the HFile the link refers to doesn't get
> {{unbuffer()}} called on it. This can cause an open Socket to hang around, as
> described in HBASE-9393.
> Both [~wchevreuil] and myself have run into this, each for different users.
> We think the commonality as to why these users saw this (but we haven't run
> into it on our own) is that it requires a very large snapshot to be brought
> into a new system. Big kudos to [~esteban] for his help in diagnosing this as
> well!
> If this analysis is accurate, it would affect all branches.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
