[
https://issues.apache.org/jira/browse/HDFS-14806?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16923707#comment-16923707
]
Chen Liang commented on HDFS-14806:
-----------------------------------
This part I am aware. Having a loop with multiple RPC calls is obvious more
overhead than one RPC call. My initial idea was based on the assumption to
continue using RPC calls here, in which case it is bug to only make that one
RPC call, and in which case the overhead of multiple calls is inevitable. But I
agree that a better fix is probably just don't use RPC call mechanism here.
The part I was confused was the "load all of the edits". I was thinking on the
"all" part because no matter what the NN needs to know about all edits. But
later I realized you probably meant to say the "load" part in that we should
not be actually loading them...Anyway, I think we are the same page now.
> 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]
