[jira] [Commented] (HADOOP-12697) IPC retry policies should recognise that SASL auth failures are unrecoverable
[ https://issues.apache.org/jira/browse/HADOOP-12697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15134330#comment-15134330 ] Steve Loughran commented on HADOOP-12697: - looks like a duplicate of HADOOP-10036. Certainly filed by same reporter, who should have checked first. > IPC retry policies should recognise that SASL auth failures are unrecoverable > - > > Key: HADOOP-12697 > URL: https://issues.apache.org/jira/browse/HADOOP-12697 > Project: Hadoop Common > Issue Type: Bug > Components: ipc >Affects Versions: 2.7.1 > Environment: Cluster with kerberos on and client not calling with the > right credentials >Reporter: Steve Loughran >Priority: Minor > > SLIDER-1050 shows that if you don't have the right kerberos settings, the > Yarn client IPC channel blocks retrying to the talk to the RM, retrying > repeatedly > {noformat} > 2016-01-07 02:50:45,111 [main] WARN ipc.Client - Exception encountered while > connecting to the server : > javax.security.sasl.SaslException: GSS initiate failed [Caused by > GSSException: > No valid credentials provided (Mechanism level: Failed to find any Kerberos > tgt)] > {noformat} > SASL exceptions need to be recognised as irreconcilable authentication > failures, rather than generic IOEs that might go away if you retry -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HADOOP-12697) IPC retry policies should recognise that SASL auth failures are unrecoverable
[ https://issues.apache.org/jira/browse/HADOOP-12697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15089717#comment-15089717 ] Sergey Shelukhin commented on HADOOP-12697: --- The retries in this case were per-exception (see YARN RMProxy), but they don't specify SaslException/GSSException anywhere. I am not sure if this is an issue with retry policy setup in YARN, or in ipc.Client. > IPC retry policies should recognise that SASL auth failures are unrecoverable > - > > Key: HADOOP-12697 > URL: https://issues.apache.org/jira/browse/HADOOP-12697 > Project: Hadoop Common > Issue Type: Bug > Components: ipc >Affects Versions: 2.7.1 > Environment: Cluster with kerberos on and client not calling with the > right credentials >Reporter: Steve Loughran >Priority: Minor > > SLIDER-1050 shows that if you don't have the right kerberos settings, the > Yarn client IPC channel blocks retrying to the talk to the RM, retrying > repeatedly > {noformat} > 2016-01-07 02:50:45,111 [main] WARN ipc.Client - Exception encountered while > connecting to the server : > javax.security.sasl.SaslException: GSS initiate failed [Caused by > GSSException: > No valid credentials provided (Mechanism level: Failed to find any Kerberos > tgt)] > {noformat} > SASL exceptions need to be recognised as irreconcilable authentication > failures, rather than generic IOEs that might go away if you retry -- This message was sent by Atlassian JIRA (v6.3.4#6332)