[ https://issues.apache.org/jira/browse/HDFS-14017?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16673712#comment-16673712 ]
Chen Liang commented on HDFS-14017: ----------------------------------- Thanks for the clarification [~xkrogen]. bq. when the active changes, how will it ever start using the new active? My thought was that, just like how current IPFailover works now, it simply assumes the VIP is pointing to ANN, changing the VIP to NN mapping happens outside of HDFS. So I was thinking of the same approach, in IPFailover, just assume the VIP is the savior when failure happens (i.e. let failoverproxy variable point to VIP always), and don't bother to figure out what exactly it is pointing to. Although the current patch may not doing this correctly/have this at all. bq. we would need to introduce additional VIPs for observers, and it's not clear to me if this makes sense Can't agree more, exactly the struggle I had! The intuition of IPFailover is to rely on VIPs and don't bother figure out the physical addresses, but when there are multiple NNs, with some being observer and some being active. I was not sure how that is going to look like. So decided to go (for now at least) IPFailover sending request to discover physical nodes by itself. > ObserverReadProxyProviderWithIPFailover should work with HA configuration > ------------------------------------------------------------------------- > > Key: HDFS-14017 > URL: https://issues.apache.org/jira/browse/HDFS-14017 > Project: Hadoop HDFS > Issue Type: Sub-task > Reporter: Chen Liang > Assignee: Chen Liang > Priority: Major > Attachments: HDFS-14017-HDFS-12943.001.patch, > HDFS-14017-HDFS-12943.002.patch > > > Currently {{ObserverReadProxyProviderWithIPFailover}} extends > {{ObserverReadProxyProvider}}, and the only difference is changing the proxy > factory to use {{IPFailoverProxyProvider}}. However this is not enough > because when calling constructor of {{ObserverReadProxyProvider}} in > super(...), the follow line: > {code:java} > nameNodeProxies = getProxyAddresses(uri, > HdfsClientConfigKeys.DFS_NAMENODE_RPC_ADDRESS_KEY); > {code} > will try to resolve the all configured NN addresses to do configured > failover. But in the case of IPFailover, this does not really apply. > > A second issue closely related is about delegation token. For example, in > current IPFailover setup, say we have a virtual host nn.xyz.com, which points > to either of two physical nodes nn1.xyz.com or nn2.xyz.com. In current HDFS, > there is always only one DT being exchanged, which has hostname nn.xyz.com. > Server only issues this DT, and client only knows the host nn.xyz.com, so all > is good. But in Observer read, even with IPFailover, the client will no > longer contacting nn.xyz.com, but will actively reaching to nn1.xyz.com and > nn2.xyz.com. During this process, current code will look for DT associated > with hostname nn1.xyz.com or nn2.xyz.com, which is different from the DT > given by NN. causing Token authentication to fail. This happens in > {{AbstractDelegationTokenSelector#selectToken}}. New IPFailover proxy > provider will need to resolve this as well. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org