[
https://issues.apache.org/jira/browse/HDFS-14806?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16922806#comment-16922806
]
Chen Liang commented on HDFS-14806:
-----------------------------------
Thanks for taking a look [~xkrogen].
bq. Calling the RPC mechanism in a loop will load all of the edits,
I didn't get this part, by "all" do you mean the specified range? I think what
I was doing there was that, instead of calling selectInputStreams on
[firstTxIdInLogs, curTxIdOnOtherNode] as it is now, I'm breaking it into
multiple calls (if necessary) of selectInputStreams on [firstTxIdInLogs,
firstTxIdInLogs+X], [firstTxIdInLogs+X+1, firstTxIdInLogs+2X],...until
[firstTxIdInLogs + nX + 1, curTxIdOnOtherNode]. Based on that it is covering
the same range, I think this loop is not loading more than what the current
code is loading. Please correct me if I'm wrong.
Or, are you suggesting even regardless of the max number of txnid, the current
code here is not efficient by calling RPC (which actually loads edits) while it
only needs to check for existences for gap checking? If this is what you meant,
I agree that using RPC here seems unnecessary and we may want to just disable
it.
> Bootstrap standby may fail if used in-progress tailing
> ------------------------------------------------------
>
> Key: HDFS-14806
> URL: https://issues.apache.org/jira/browse/HDFS-14806
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: namenode
> Affects Versions: 3.3.0
> Reporter: Chen Liang
> Assignee: Chen Liang
> Priority: Major
> Attachments: HDFS-14806.001.patch, HDFS-14806.002.patch
>
>
> One issue we went across was that if in-progress tailing is enabled,
> bootstrap standby could fail.
> When in-progress tailing is enabled, Bootstrap uses the RPC mechanism to get
> edits. There is a config {{dfs.ha.tail-edits.qjm.rpc.max-txns}} that sets an
> upper bound on how many txnid can be included in one RPC call. The default is
> 5000. Meaning bootstraping NN (say NN1) can only pull at most 5000 edits from
> JN. However, as part of bootstrap, NN1 queries another NN (say NN2) for NN2's
> current transactionID, NN2 may return a state that is > 5000 txnid from NN1's
> current image. But NN1 can only see 5000 more txnid from JNs. At this point
> NN1 goes panic, because txnid retuned by JNs is behind NN2's returned state,
> bootstrap then fail.
> Essentially, bootstrap standby can fail if both of two following conditions
> are met:
> # in-progress tailing is enabled AND
> # the boostraping NN is too far (>5000 txid) behind
> Increasing the value of {{dfs.ha.tail-edits.qjm.rpc.max-txns}} to some super
> large value allowed bootstrap to continue. But this is hardly the ideal
> solution.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
