[
https://issues.apache.org/jira/browse/HBASE-2856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12989399#comment-12989399
]
ryan rawson commented on HBASE-2856:
------------------------------------
Don't we already have this? The comparator uses the max_seq_id to break ties
between KVs...
The primary issue is that we need to know which KVs are 'committed' and which
are still being created in progress. Right now we have a problem whereby the
scanner stack gets a little wonky about how it handles partial next()s. By
moving the memstoreTS pruning up to the HRegion scanner level, and working on
entire rows at a time, this might mitigate most of the problem actually. This
might get ugly with family only flushes, since in theory you might end up with
a row that is not completely written but is in memstore & hfile at the same
time. Given that the scope of a RWCC "transaction" is only memstore insert, I'm
not sure how that would happen. It's possible we could prevent it from becoming
a problem with judicious use of the updateLock in HRegion though.
For example, by grabbing the updateLock.writeLock().lock() during the switch
over, or the flush, we could ensure that all the pending writes are now
complete, then do the switch out, then we'd never have a situation where a half
committed write is in memstore & hfile at the same time.
> TestAcidGuarantee broken on trunk
> ----------------------------------
>
> Key: HBASE-2856
> URL: https://issues.apache.org/jira/browse/HBASE-2856
> Project: HBase
> Issue Type: Bug
> Affects Versions: 0.89.20100621
> Reporter: ryan rawson
> Assignee: stack
> Priority: Blocker
> Fix For: 0.92.0
>
> Attachments: 2856-v2.txt, 2856-v3.txt, 2856-v4.txt, 2856-v5.txt,
> acid.txt
>
>
> TestAcidGuarantee has a test whereby it attempts to read a number of columns
> from a row, and every so often the first column of N is different, when it
> should be the same. This is a bug deep inside the scanner whereby the first
> peek() of a row is done at time T then the rest of the read is done at T+1
> after a flush, thus the memstoreTS data is lost, and previously 'uncommitted'
> data becomes committed and flushed to disk.
> One possible solution is to introduce the memstoreTS (or similarly equivalent
> value) to the HFile thus allowing us to preserve read consistency past
> flushes. Another solution involves fixing the scanners so that peek() is not
> destructive (and thus might return different things at different times alas).
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
