[
https://issues.apache.org/jira/browse/HADOOP-7472?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13072006#comment-13072006
]
Suresh Srinivas commented on HADOOP-7472:
-----------------------------------------
bq. We are working on a new approach, which will address both 1 and 2.
Can you add more details.
bq. They don't. It's that they create an InetSocketAddress and the lower layers
have no way of knowing what it was originally instantiated with. This is a
headache when dealing with tokens.
You could consider host name from the InetSocketAddress as the destination
identification. I am not sure what you mean by "originally instantiated with"
and why treating InetSocketAddress as carrier of host name information will not
work.
bq. This is a headache when dealing with tokens.
I am not sure what the headache. I will ask Jitendra to comment on this from
Security perspective.
bq. We thought about this. Darren is working on the token renewal problem and
we found out we can have a common solution. One way was to do what you
mentioned. But decided to keep it as is but use createUnresolved() to create an
InetSocketAddress, so that we know what was used to instantiate it. If the user
slapped in an IP address to begin with, we won't handle it. (I think it was
indistinguishable before)
Does this mean, you will have different implementation later? Like I said, we
could treat InetSocketAddress as carrier of hostname and not attach any other
semantics to it. This should be fine because InetSocketAddress is what is
passed. And the name(URL) is also resolved to InetSocketAddress.
bq. The token will have whatever the user used (IP or name) in the beginning
and in case of using name, the key to the token cache won't change even with
addr changes. So the delegation token should continue to work.
Jitendra, any comments on this? My thought is, if you wrap InetSocketAddress,
appropriate key such as name or host name from InetSocketAddress could be used.
> RPC client should deal with the IP address changes
> --------------------------------------------------
>
> Key: HADOOP-7472
> URL: https://issues.apache.org/jira/browse/HADOOP-7472
> Project: Hadoop Common
> Issue Type: Improvement
> Components: ipc
> Affects Versions: 0.20.205.0
> Reporter: Kihwal Lee
> Assignee: Kihwal Lee
> Priority: Minor
> Fix For: 0.20.205.0
>
> Attachments: addr_change_dfs-1.patch.txt, addr_change_dfs.patch.txt
>
>
> The current RPC client implementation and the client-side callers assume that
> the hostname-address mappings of servers never change. The resolved address
> is stored in an immutable InetSocketAddress object above/outside RPC, and the
> reconnect logic in the RPC Connection implementation also trusts the resolved
> address that was passed down.
> If the NN suffers a failure that requires migration, it may be started on a
> different node with a different IP address. In this case, even if the
> name-address mapping is updated in DNS, the cluster is stuck trying old
> address until the whole cluster is restarted.
> The RPC client-side should detect this situation and exit or try to recover.
> Updating ConnectionId within the Client implementation may get the system
> work for the moment, there always is a risk of the cached address:port become
> connectable again unintentionally. The real solution will be notifying upper
> layer of the address change so that they can re-resolve and retry or
> re-architecture the system as discussed in HDFS-34.
> For 0.20 lines, some type of compromise may be acceptable. For example, raise
> a custom exception for some well-defined high-impact upper layer to do
> re-resolve/retry, while other will have to restart. For TRUNK, the HA work
> will most likely determine what needs to be done. So this Jira won't cover
> the solutions for TRUNK.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
