[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-07 Thread Todd Lipcon (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13249181#comment-13249181 ] Todd Lipcon commented on HDFS-3192: --- The state diagram is included in the design doc

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-06 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13249099#comment-13249099 ] Hari Mankude commented on HDFS-3192: Would it be possible to post state transition

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Suresh Srinivas (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246059#comment-13246059 ] Suresh Srinivas commented on HDFS-3192: --- Hari, agree with Aaron that this should not

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246661#comment-13246661 ] Hari Mankude commented on HDFS-3192: bq.Why? I think it's an advantage that the FC may

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Todd Lipcon (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246671#comment-13246671 ] Todd Lipcon commented on HDFS-3192: --- Why add multiple stonith paths, given we need

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246687#comment-13246687 ] Hari Mankude commented on HDFS-3192: bq. Why add multiple stonith paths, given we need

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Todd Lipcon (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246718#comment-13246718 ] Todd Lipcon commented on HDFS-3192: --- bq. I thought we are not going to have external

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246752#comment-13246752 ] Hari Mankude commented on HDFS-3192: bq. bq. Why is the behaviour different from what

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Todd Lipcon (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246778#comment-13246778 ] Todd Lipcon commented on HDFS-3192: --- I think there is confusion here over the terminology

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246805#comment-13246805 ] Hari Mankude commented on HDFS-3192: Excellent comment regarding quorum and active

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246806#comment-13246806 ] Hari Mankude commented on HDFS-3192: bq.Excellent comment regarding quorum and active

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Todd Lipcon (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246818#comment-13246818 ] Todd Lipcon commented on HDFS-3192: --- bq. So in step #6, irrespective of when ZKFC1 gets

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246852#comment-13246852 ] Hari Mankude commented on HDFS-3192: bq.Can you explain why it has to restart, instead

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Todd Lipcon (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246865#comment-13246865 ] Todd Lipcon commented on HDFS-3192: --- bq. Are you suggesting that ZKFC1 does

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Todd Lipcon (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246868#comment-13246868 ] Todd Lipcon commented on HDFS-3192: --- Maybe the confusion is about readers? The one hole

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-04 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13246949#comment-13246949 ] Hari Mankude commented on HDFS-3192: bq.2a) If the local node is still accessible, then

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-03 Thread Hari Mankude (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13245889#comment-13245889 ] Hari Mankude commented on HDFS-3192: FC via healthmonitor will periodically poll the

[jira] [Commented] (HDFS-3192) Active NN should exit when it has not received a getServiceStatus() rpc from ZKFC for timeout secs

2012-04-03 Thread Todd Lipcon (Commented) (JIRA)
[ https://issues.apache.org/jira/browse/HDFS-3192?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13245895#comment-13245895 ] Todd Lipcon commented on HDFS-3192: --- Why? I think it's an _advantage_ that the FC may