lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Erick Erickson (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (LUCENE-7976) Make TieredMergePolicy respect maxSegmentSizeMB and allow singleton merges of very large segments
Date Sun, 22 Apr 2018 00:05:00 GMT

    [ https://issues.apache.org/jira/browse/LUCENE-7976?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16447046#comment-16447046
] 

Erick Erickson commented on LUCENE-7976:
----------------------------------------

[~mikemccand] Thanks for looking!

About removing  {{@lucene.experimental}}, yes that was deliberate, TMP has been around for
a very long time and it seemed to me that it's now mainstream. I have no problem with putting
it back. Let me know if that's your preference. Is putting it back for back-compat? Well,
actually so we don't _have_ to maintain back-compat?

bq. Can we do this change in two parts? First part is the nice refactoring to have all the
methods share a common scoring loop, which should show no behavior change I think?

Maybe I got the block quote thing right this time, thanks!

What's the purpose here? Mechanically it's simple and I'll be glad to do it, I'd just like
to know what the goal is. My guess is so we can have a clear distinction between changes in
behavior in NATURAL indexing and refactoring.

When you say "no change in behavior" you were referring to NATURAL merging, correct? Not FORCE_MERGE
or FORCE_MERGE_DELETES. Those will behave quite differently.

bq. can you name it NATURAL, FORCE_MERGE and FORCE_MERGE_DELETES?

done.

{{quote}}
Hmm what is the note/NOTD? Can you change to As of Lucene 7.4
{{quote}}
What can I say? I spend 99% of my life in Solr, _everything_ is Solr, right? As for rest,
typos late at night.

Done.

Finally, can you comment on this nocommit?
{{quote}}
Should you be using writer.numDeletesToMerge rather than the info.getDelDocs other places
{{quote}}

I see both of these in the code, and writer.numDeletesToMerge seems considerably more expensive.
Is there a reason to prefer one over the other?

Thanks again!

> Make TieredMergePolicy respect maxSegmentSizeMB and allow singleton merges of very large
segments
> -------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-7976
>                 URL: https://issues.apache.org/jira/browse/LUCENE-7976
>             Project: Lucene - Core
>          Issue Type: Improvement
>            Reporter: Erick Erickson
>            Assignee: Erick Erickson
>            Priority: Major
>         Attachments: LUCENE-7976.patch, LUCENE-7976.patch, LUCENE-7976.patch, LUCENE-7976.patch
>
>
> We're seeing situations "in the wild" where there are very large indexes (on disk) handled
quite easily in a single Lucene index. This is particularly true as features like docValues
move data into MMapDirectory space. The current TMP algorithm allows on the order of 50% deleted
documents as per a dev list conversation with Mike McCandless (and his blog here:  https://www.elastic.co/blog/lucenes-handling-of-deleted-documents).
> Especially in the current era of very large indexes in aggregate, (think many TB) solutions
like "you need to distribute your collection over more shards" become very costly. Additionally,
the tempting "optimize" button exacerbates the issue since once you form, say, a 100G segment
(by optimizing/forceMerging) it is not eligible for merging until 97.5G of the docs in it
are deleted (current default 5G max segment size).
> The proposal here would be to add a new parameter to TMP, something like <maxAllowedPctDeletedInBigSegments>
(no, that's not serious name, suggestions welcome) which would default to 100 (or the same
behavior we have now).
> So if I set this parameter to, say, 20%, and the max segment size stays at 5G, the following
would happen when segments were selected for merging:
> > any segment with > 20% deleted documents would be merged or rewritten NO MATTER
HOW LARGE. There are two cases,
> >> the segment has < 5G "live" docs. In that case it would be merged with smaller
segments to bring the resulting segment up to 5G. If no smaller segments exist, it would just
be rewritten
> >> The segment has > 5G "live" docs (the result of a forceMerge or optimize).
It would be rewritten into a single segment removing all deleted docs no matter how big it
is to start. The 100G example above would be rewritten to an 80G segment for instance.
> Of course this would lead to potentially much more I/O which is why the default would
be the same behavior we see now. As it stands now, though, there's no way to recover from
an optimize/forceMerge except to re-index from scratch. We routinely see 200G-300G Lucene
indexes at this point "in the wild" with 10s of  shards replicated 3 or more times. And that
doesn't even include having these over HDFS.
> Alternatives welcome! Something like the above seems minimally invasive. A new merge
policy is certainly an alternative.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message