[
https://issues.apache.org/jira/browse/HBASE-10227?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13871201#comment-13871201
]
Lars Hofhansl commented on HBASE-10227:
---------------------------------------
I have not followed all the discussion here. The key points seem to be:
# we'll always keep the read point (never set it 0)
# that in turns means we always have to store and decode it upon read
So for each KV we'll store an extra vlong. Since we don't set it to null we'll
likely need at least 4 bytes. In that case we should probably not store as a
vlong, but just an 8 byte long.
For small KeyValue that would probably make a noticeable performance impact,
but probably acceptable if needed for correctness.
What I do not understand what we're trying to solve. As described in the
description we do not allow a scanner before splitting is finished, so this is
not currently a problem. How would we allow opening a scanner before we open
the region?
> When a region is opened, its mvcc isn't correctly recovered when there are
> split hlogs to replay
> ------------------------------------------------------------------------------------------------
>
> Key: HBASE-10227
> URL: https://issues.apache.org/jira/browse/HBASE-10227
> Project: HBase
> Issue Type: Bug
> Components: regionserver
> Reporter: Feng Honghua
> Assignee: Feng Honghua
> Attachments: HBASE-10227-trunk_v0.patch
>
>
> When opening a region, all stores are examined to get the max MemstoreTS and
> it's used as the initial mvcc for the region, and then split hlogs are
> replayed. In fact the edits in split hlogs have kvs with greater mvcc than
> all MemstoreTS in all store files, but replaying them don't increment the
> mvcc according at all. From an overall perspective this mvcc recovering is
> 'logically' incorrect/incomplete.
> Why currently it doesn't incur problem is because no active scanners exists
> and no new scanners can be created before the region opening completes, so
> the mvcc of all kvs in the resulted hfiles from hlog replaying can be safely
> set to zero. They are just treated as kvs put 'earlier' than the ones in
> HFiles with mvcc greater than zero(say 'earlier' since they have mvcc less
> than the ones with non-zero mvcc, but in fact they are put 'later'), and
> without any incorrect impact just because during region opening there are no
> active scanners existing / created.
> This bug is just in 'logic' sense for the time being, but if later on we need
> to survive mvcc in the region's whole logic lifecycle(across regionservers)
> and never set them to zero, this bug needs to be fixed first.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
