[
https://issues.apache.org/jira/browse/HADOOP-6726?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12861132#action_12861132
]
Christian Kunz commented on HADOOP-6726:
----------------------------------------
I completely agree. A client's retry behavior should not be hard-coded into a
library. Imagine a high number of clients overwhelming the namenode with
requests -- currently there is no easy way for clients to implement a decent
backoff scheme such that the namenode can recover.
> can't control maxRetries in case of SocketTimeoutException
> ----------------------------------------------------------
>
> Key: HADOOP-6726
> URL: https://issues.apache.org/jira/browse/HADOOP-6726
> Project: Hadoop Common
> Issue Type: Improvement
> Components: ipc
> Affects Versions: 0.20.2
> Reporter: Johannes Zillmann
>
> One can set _ipc.client.connect.max.retries_ for
> _org.apache.hadoop.ipc.Client_.
> This comes to effect on IOExceptions but not on SocketTimeoutException.
> Client$Connection:307:
> {code:java}
> } catch (SocketTimeoutException toe) {
> /* The max number of retries is 45,
> * which amounts to 20s*45 = 15 minutes retries.
> */
> handleConnectionFailure(timeoutFailures++, 45, toe);
> } catch (IOException ie) {
> handleConnectionFailure(ioFailures++, maxRetries, ie);
> }
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
