[
https://issues.apache.org/jira/browse/HBASE-15400?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15191804#comment-15191804
]
Clara Xiong commented on HBASE-15400:
-------------------------------------
[~Apache9]Coming back to the discussion on assigning distinct seq id on the
output files, we cannot safely do it unless we have enough space between 0 and
HRegion.sequenceId. This is not only for Date Tiered Compaction but also for
Stripe Compaction. Going into negative breaks the assumption of sequence id
being non negative except for bulk load files with assigning seg id turned off.
One way is to bump up HRegion.sequenceId to max output count we need for
compaction every time we request compaction so that the future flushes will
carry larger seq id when we are figuring out whether we need to compact.
Another way is to force every flush/bulk load to set HRegion.sequenceId to
max(HRegion.sequenceId + 1, max output count). I prefer the second way. But
either way requires quite some major changes and the benefit is limited. We
know correctness will be guaranteed among output files for stripe and date
tiered compaction because there is either no dup key or no dup timestamp among
different files. We may want to put this work into a separate jira and address
it when it is required.
> Use DateTieredCompactor for Date Tiered Compaction
> --------------------------------------------------
>
> Key: HBASE-15400
> URL: https://issues.apache.org/jira/browse/HBASE-15400
> Project: HBase
> Issue Type: Sub-task
> Components: Compaction
> Reporter: Clara Xiong
> Assignee: Clara Xiong
> Fix For: 2.0.0
>
> Attachments: HBASE-15400.patch
>
>
> When we compact, we can output multiple files along the current window
> boundaries. There are two use cases:
> 1. Major compaction: We want to output date tiered store files with data
> older than max age archived in trunks of the window size on the higher tier.
> 2. Bulk load files and the old file generated by major compaction before
> upgrading to DTCP.
> Pros:
> 1. Restore locality, process versioning, updates and deletes while
> maintaining the tiered layout.
> 2. The best way to fix a skewed layout.
>
> This work is based on a prototype of DateTieredCompactor from HBASE-15389 and
> focused on the part to meet needs for these two use cases while supporting
> others. I have to call out a few design decisions:
> 1. We only want to output the files along all windows for major compaction.
> And we want to output multiple files older than max age in the trunks of the
> maximum tier window size determined by base window size, windows per tier and
> max age.
> 2. For minor compaction, we don't want to output too many files, which will
> remain around because of current restriction of contiguous compaction by seq
> id. I will only output two files if all the files in the windows are being
> combined, one for the data within window and the other for the out-of-window
> tail. If there is any file in the window excluded from compaction, only one
> file will be output from compaction. When the windows are promoted, the
> situation of out of order data will gradually improve. For the incoming
> window, we need to accommodate the case with user-specified future data.
> 3. We have to pass the boundaries with the list of store file as a complete
> time snapshot instead of two separate calls because window layout is
> determined by the time the computation is called. So we will need new type of
> compaction request.
> 4. Since we will assign the same seq id for all output files, we need to sort
> by maxTimestamp subsequently. Right now all compaction policy gets the files
> sorted for StoreFileManager which sorts by seq id and other criteria. I will
> use this order for DTCP only, to avoid impacting other compaction policies.
> 5. We need some cleanup of current design of StoreEngine and CompactionPolicy.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
