Github user harishreedharan commented on the pull request:
https://github.com/apache/spark/pull/7255#issuecomment-119861362
@tgravescs - So this fixes this one issue with yarn-client mode, but there
is one issue that actually has always been the case with this one in
yarn-client. Since we don't re-login, it looks like we can't actually write to
HDFS from user code. My assumption was that the HDFS client API takes care of
this, but I get this on the application logs, while the executors continued to
write without issues:
```
org.apache.hadoop.ipc.RemoteException(org.apache.hadoop.security.token.SecretManager$InvalidToken):
token (HDFS_DELEGATION_TOKEN token 66 for hari) is expired
at org.apache.hadoop.ipc.Client.call(Client.java:1468)
at org.apache.hadoop.ipc.Client.call(Client.java:1399)
at
org.apache.hadoop.ipc.ProtobufRpcEngine$Invoker.invoke(ProtobufRpcEngine.java:232)
at com.sun.proxy.$Proxy12.renewLease(Unknown Source)
at
org.apache.hadoop.hdfs.protocolPB.ClientNamenodeProtocolTranslatorPB.renewLease(ClientNamenodeProtocolTranslatorPB.java:571)
at sun.reflect.GeneratedMethodAccessor19.invoke(Unknown Source)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:606)
at
org.apache.hadoop.io.retry.RetryInvocationHandler.invokeMethod(RetryInvocationHandler.java:187)
at
org.apache.hadoop.io.retry.RetryInvocationHandler.invoke(RetryInvocationHandler.java:102)
at com.sun.proxy.$Proxy13.renewLease(Unknown Source)
at org.apache.hadoop.hdfs.DFSClient.renewLease(DFSClient.java:878)
at org.apache.hadoop.hdfs.LeaseRenewer.renew(LeaseRenewer.java:417)
at org.apache.hadoop.hdfs.LeaseRenewer.run(LeaseRenewer.java:442)
at org.apache.hadoop.hdfs.LeaseRenewer.access$700(LeaseRenewer.java:71)
at org.apache.hadoop.hdfs.LeaseRenewer$1.run(LeaseRenewer.java:298)
at java.lang.Thread.run(Thread.java:745)
```
So either we can doc this and ask the user to `kinit` or somehow trigger
re-logins in `SparkSubmit` (I am already not a fan of the current change in
`SparkSubmit` though). I am fine with either - what do you suggest?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
