[ https://issues.apache.org/jira/browse/LUCENE-2324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12984738#action_12984738 ]
Michael McCandless commented on LUCENE-2324: -------------------------------------------- bq. If we implemented sequence-id deletes using a short[], then we're only increasing the RAM usage by 16 times, though we then do not need to clone which can generate excessive garbage (in a high flush [N]RT enviro). How would we handle wraparound (in a concurrent way)? Also, 16 fold increase in RAM usage is not cheap! I think, instead, we should recycle the bit vectors? See, what we do now is really quite silly: we drop the BV, let GC recycle it, allocate a new BV (same size), copy in nearly the same bits that we just discarded, set a few more bits. If instead we had a pool that'd hold recently freed BVs (for a given segment), but, also tracked their "state" ie what "delete gen" they were at, and then when we need a new BV for that same segment, we pull a free one, catch it up (replay the deletes that had arrived since it was created), and use it, that's very fast? Ie the cost is about as good as it can be -- incremental to the number of deletes actually recorded. And the added RAM is "a few bits" per doc, where exactly how many "a few" is is decided by your app, ie, how many in-flight readers it "tends" to keep open. > Per thread DocumentsWriters that write their own private segments > ----------------------------------------------------------------- > > Key: LUCENE-2324 > URL: https://issues.apache.org/jira/browse/LUCENE-2324 > Project: Lucene - Java > Issue Type: Improvement > Components: Index > Reporter: Michael Busch > Assignee: Michael Busch > Priority: Minor > Fix For: Realtime Branch > > Attachments: LUCENE-2324-SMALL.patch, LUCENE-2324-SMALL.patch, > LUCENE-2324-SMALL.patch, LUCENE-2324-SMALL.patch, LUCENE-2324-SMALL.patch, > LUCENE-2324.patch, LUCENE-2324.patch, LUCENE-2324.patch, lucene-2324.patch, > lucene-2324.patch, LUCENE-2324.patch, test.out, test.out, test.out, test.out > > > See LUCENE-2293 for motivation and more details. > I'm copying here Mike's summary he posted on 2293: > Change the approach for how we buffer in RAM to a more isolated > approach, whereby IW has N fully independent RAM segments > in-process and when a doc needs to be indexed it's added to one of > them. Each segment would also write its own doc stores and > "normal" segment merging (not the inefficient merge we now do on > flush) would merge them. This should be a good simplification in > the chain (eg maybe we can remove the *PerThread classes). The > segments can flush independently, letting us make much better > concurrent use of IO & CPU. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org