OK thanks for the updates. Yes, we are on the hunt now ;) Something nasty is lurking...
Weird. Have you run CheckIndex? Can you do a binary search to locate the term(s) that's causing it? It's great you see 10% speedup in searching overall (excluding these ones...)! Mike On Wed, Jul 15, 2009 at 1:49 PM, eks dev<eks...@yahoo.co.uk> wrote: > > > 1. pls forget minNumberShould match, it is NOT set on this particular query > (minNumberShouldMatch is determined dynamically, depending on semantics of > user query... sometimes triggers, sometimes not...). > This Exact Query here causes search to take longer than 180 Seconds with > allowDocsOutOfOrder = true, and less than 70mS with false. Repeatable?!? No > gc() effects involved... on 2.4 it does not happen, it works fine with both > true/false for allowDocsOutOfOrder > > 2. re your test, That is exactly what makes me wonder, we also see average > performance almost 10% better on 2.9 (even on this index when we exclude > these stuck searches), but on this particular index our customer's QA > managed to find these "stuck requests". > > 3. If I change tokens involved, in exactly same-structured Query, it runs > fine => The problem is somehow term-defendant (bah!) > > Please understand that I do not have direct access to this index and it makes > debug cycles slightly longer. Typically I give them some jar-s and they run > it ans send me logs back... Sorry for inaccuracies in description, but I am > sure there is a problem in lucene... We tried it with Luke as well, freshly > built index, we see exactly the same behavior (no bugs in our app that could > cause it, except maybe wrong lucene usage somewhere) > > > Hard, but please stay with me, we will fix one ugly bug :) > > > > > > > > ----- Original Message ---- >> From: Michael McCandless <luc...@mikemccandless.com> >> To: java-user@lucene.apache.org >> Sent: Wednesday, 15 July, 2009 19:27:24 >> Subject: Re: speed of BooleanQueries on 2.9 >> >> But, that query can't accept a minNumberShouldMatch -- are you really >> setting that? (You get 0 results if you set it, because the top >> boolean query has a single required clause). Maybe you set it only on >> the inner large OR-query? (But then I don't see the ~2 on that inner >> clause). >> >> I've tested a 21 term OR query, with allowDocsOutOfOrder true, >> numHits=200 on a Wikpedia index that matches 10M docs and I'm seeing >> the same perf on trunk & 2.4. >> >> Mike >> >> On Wed, Jul 15, 2009 at 11:41 AM, eks devwrote: >> > >> > sorry for confusion, here is exact query that runs forever with >> setAllowDocsOutOfOrder: >> > You see it on stack trace taken while "stuck" >> o.a.l.search.TopScoreDocCollector$OutOfOrderTopScoreDocCollector.collect(UnknownSource) >> > >> > >> > Query: +(((NAME:maria NAME:marae^0.25171682 NAME:marai^0.2365632 >> NAME:marao^0.2365632 NAME:marau^0.2365632 NAME:marea^0.2834352 >> NAME:marei^0.25171682 NAME:mareo^0.25171682 NAME:mareu^0.25171682 >> NAME:marie^0.28577283 NAME:marieh^0.2451648 NAME:mariha^0.2583552 >> NAME:mariu^0.27189124 NAME:marja^0.2834352 NAME:marje^0.2673408 >> NAME:marji^0.25171682 NAME:marjo^0.25171682 NAME:marju^0.25171682 >> NAME:marla^0.2673408 NAME:marle^0.25171682 NAME:marli^0.2365632 >> NAME:marlo^0.2365632 NAME:maroa^0.2673408 NAME:maroe^0.25171682 >> NAME:maroi^0.2365632 NAME:marou^0.2365632 NAME:marua^0.2673408 >> NAME:marue^0.25171682 NAME:marui^0.2365632 NAME:maruo^0.2365632 >> NAME:marye^0.2673408 NAME:maryi^0.25171682 NAME:maryo^0.25171682 >> NAME:meria^0.2787888 NAME:miria^0.25835523 NAME:moria^0.25835523 >> NAME:muria^0.25835523 NAME:naria^0.27648002 NAME:narie^0.25392002 >> NAME:neria^0.25392002) (NAME:piekarski NAME:bekarski^0.19200002 >> NAME:beugarski^0.20281483 NAME:blacharski^0.19200002 >> > NAME:lekarski^0.19200002 NAME:pecarski^0.21294187 >> > NAME:peikarski^0.27648002 >> NAME:pekarska^0.20172001 NAME:pekarski^0.22446752 NAME:pekarskj^0.21294187 >> NAME:pekarsky^0.21294187 NAME:pickarske^0.21168004 NAME:pickarski^0.22073482 >> NAME:piekalski^0.23941332 NAME:piekanski^0.23941332 NAME:piekaraka^0.22533335 >> NAME:piekarsci^0.29205337 NAME:piekarska^0.28421336 >> NAME:piekarskie^0.25392002 >> NAME:piekarsky^0.29205337 NAME:piekarzcyk^0.23232001 >> NAME:piekarzki^0.29205337 >> NAME:piekaski^0.24843001 NAME:piekavska^0.22533335 NAME:piekorski^0.28421336 >> NAME:pielarski^0.22997928 NAME:pierarski^0.22997928 >> NAME:pierkarski^0.24661335 >> NAME:piesarski^0.22997928 NAME:pietarski^0.22997928 >> NAME:pietkarski^0.24661335 >> NAME:pikarski^0.23232001 NAME:piowarski^0.20281483 NAME:pirkarski^0.22073482 >> NAME:plocharski^0.21168004 NAME:pokarski^0.20172001 >> NAME:polikarski^0.20172001 >> NAME:pukarski^0.20172001 NAME:pyekarska^0.26508 >> NAME:siekarski^0.20281483))^2.0) >> > >> > >> > >> > >> > >> > ----- Original Message ---- >> >> From: Michael McCandless >> >> To: java-user@lucene.apache.org >> >> Sent: Wednesday, 15 July, 2009 17:16:23 >> >> Subject: Re: speed of BooleanQueries on 2.9 >> >> >> >> So now I'm confused. Since your query has required (+) clauses, the >> >> setAllowDocsOutOfOrder should have no effect, on either 2.4 or trunk. >> >> >> >> BooleanQuery only uses BooleanScorer when there are no required terms, >> >> and allowDocsOutOfOrder is true. So I can't explain why you see this >> >> setting changing anything on this query... >> >> >> >> Mike >> >> >> >> On Tue, Jul 14, 2009 at 7:04 PM, eks devwrote: >> >> > >> >> > I do not know exactly why, but >> >> > when I BooleanQuery.setAllowDocsOutOfOrder(true); I have the problem, >> >> > but >> with >> >> setAllowDocsOutOfOrder(false); no problems whatsoever >> >> > >> >> > not really scientific method to find such bug, but does the job and >> >> > makes >> me >> >> happy. >> >> > >> >> > Empirical, "deprecated methods are not to be taken as thoroughly >> >> > tested, as >> >> they have short life expectancy" >> >> > >> >> > >> >> > >> >> > >> >> > >> >> > ----- Original Message ---- >> >> >> From: eks dev >> >> >> To: java-user@lucene.apache.org >> >> >> Sent: Wednesday, 15 July, 2009 0:24:43 >> >> >> Subject: Re: speed of BooleanQueries on 2.9 >> >> >> >> >> >> >> >> >> Mike, we are definitely hitting something with this one! >> >> >> >> >> >> we had report from our QA chaps that our servers got stuck (limit is on >> 180 >> >> >> Seconds Request)... We are on average 14 Requsts per second.... has >> nothing >> >> to >> >> >> do with gc() as >> >> >> we can repeat it with freshly restarted searcher. >> >> >> >> >> >> - it happens on a less than 0.1% of queries, not much of a pattern, >> >> repeatable >> >> >> on our index... >> >> >> it is always combination of two expanded tokens (we use >> >> >> minimumNooShouldMatch)... >> >> >> >> >> >> (+(t1 [up to 40 expansions]) +(t2 [up to 40 expansions of t2])) >> >> >> all tokens are with set boost, and minNumShouldMatch is set to two >> >> >> >> >> >> I cannot provide self-contained test, nor index (contains sensitive >> >> >> data >> and >> >> is >> >> >> rather big, ~5G) >> >> >> >> >> >> I can repeat this test on t1 and t2 with 40 expansions each. even if I >> take >> >> the >> >> >> most frequent tokens in collection it runs well under one second...but >> these >> >> two >> >> >> particular tokens with their "expansions" are making it run forever... >> >> >> >> >> >> and yes, if I run t1 plus expansions only, it runs super fast, the same >> for >> >> t2 >> >> >> >> >> >> java 1.4U14, tried wit 1.6U6, no changes... >> >> >> >> >> >> will report if I dig something out >> >> >> >> >> >> partial stack trace while "stuck", cpu is on max: >> >> >> >> >> >> >> >> >> org.apache.lucene.search.TopScoreDocCollector$OutOfOrderTopScoreDocCollector.collect(Unknown >> >> >> Source) >> >> >> org.apache.lucene.search.BooleanScorer.score(Unknown Source) >> >> >> org.apache.lucene.search.BooleanScorer.score(Unknown Source) >> >> >> org.apache.lucene.search.IndexSearcher.search(Unknown Source) >> >> >> org.apache.lucene.search.IndexSearcher.search(Unknown Source) >> >> >> org.apache.lucene.search.Searcher.search(Unknown Source) >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> ----- Original Message ---- >> >> >> > From: eks dev >> >> >> > To: java-user@lucene.apache.org >> >> >> > Sent: Monday, 13 July, 2009 13:28:45 >> >> >> > Subject: Re: speed of BooleanQueries on 2.9 >> >> >> > >> >> >> > Hi Mike, >> >> >> > >> >> >> > getMaxNumOfCandidates() in test was 200, Index is optimised and >> read-only >> >> >> > >> >> >> > We found (due to an error in our warm-up code, funny) that only this >> Query >> >> >> runs >> >> >> > slower on 2.9. >> >> >> > >> >> >> > A hint where to look could be that this Query cointains two, the most >> >> frequent >> >> >> >> >> >> > tokens in two particular fields >> >> >> > NAME:hans and ZIPS:berlin (index has ca 80Mio very short documents, >> >> >> > 3Mio >> >> >> unique >> >> >> > terms) >> >> >> > >> >> >> > But all of this *could be just wrong measurement*, I just could not >> spend >> >> more >> >> >> >> >> >> > time to get to the bottom of this. We moved forward as we got overall >> >> better >> >> >> > average performance (sweet 10% in average) on much bigger real query >> >> >> > log >> >> from >> >> >> > our regression test. >> >> >> > >> >> >> > Anyhow I just wanted to throw it out, maybe it triggers some >> >> >> > synapses :) >> If >> >> >> > false alarm, sorry. >> >> >> > >> >> >> > >> >> >> > >> >> >> > >> >> >> > >> >> >> > ----- Original Message ---- >> >> >> > > From: Michael McCandless >> >> >> > > To: java-user@lucene.apache.org >> >> >> > > Sent: Monday, 13 July, 2009 11:50:48 >> >> >> > > Subject: Re: speed of BooleanQueries on 2.9 >> >> >> > > >> >> >> > > This is not expected; 2.9 has had a number of changes that ought to >> >> >> > > reduce CPU cost of searching. If this holds up we definitely need >> >> >> > > to >> >> >> > > get to the root cause. >> >> >> > > >> >> >> > > Did your test exclude the warmup query for both 2.4.1 & 2.9? How >> >> >> > > many >> >> >> > > segments in the index? What is the actual value of >> >> >> > > getMaxNumOfCandidates()? If you simplify the query down (eg just >> >> >> > > do >> >> >> > > the NAME clause or the ZIPSS clause, alone) are those also 4X >> >> >> > > slower? >> >> >> > > >> >> >> > > Mike >> >> >> > > >> >> >> > > On Sun, Jul 12, 2009 at 12:53 PM, eks devwrote: >> >> >> > > > >> >> >> > > > Is it possible that the same BooleanQuery on 2.9 runs >> >> >> > > > significantly >> >> slower >> >> >> >> >> >> > > than on 2.4? >> >> >> > > > >> >> >> > > > we have some strange effects where the following query runs >> >> >> > > > approx >> >> >> 4(ouch!) >> >> >> > > times slower on 2.9, test done by 1000 times executing the same >> Query... >> >> >> But! >> >> >> > if >> >> >> > > I run test from some real Query log with mixed Queries, I get >> >> >> > > almost >> the >> >> >> same >> >> >> > > results (?!), even slightly faster on 2.9 !? >> >> >> > > > >> >> >> > > > >> >> >> > > > Query: >> >> >> > > > +((NAME:hans NAME:hahns^0.23232001 NAME:hams^0.27648002 >> >> NAME:hamz^0.25392 >> >> >> > > NAME:hanas^0.18722998 NAME:hanbs^0.18722998 NAME:hanfs^0.18722998 >> >> >> > > NAME:hangs^0.18722998 NAME:hanhs^0.24030754 NAME:hanis^0.18722998 >> >> >> > > NAME:hanjs^0.18722998 NAME:hanks^0.18722998 NAME:hanms^0.18722998 >> >> >> > > NAME:hanos^0.18722998 NAME:hanrs^0.18722998 NAME:hansb^0.20172001 >> >> >> > > NAME:hansd^0.20172001 NAME:hansf^0.20172001 NAME:hansg^0.20172001 >> >> >> > > NAME:hansi^0.20172001 NAME:hansj^0.20172001 NAME:hansk^0.20172001 >> >> >> > > NAME:hansl^0.20172001 NAME:hansn^0.20172001 NAME:hanso^0.20172001 >> >> >> > > NAME:hansp^0.20172001 NAME:hanst^0.20172001 NAME:hansu^0.20172001 >> >> >> > > NAME:hansw^0.20172001 NAME:hansy^0.20172001 NAME:hansz^0.20172001 >> >> >> > > NAME:hants^0.18722998 NAME:hanus^0.18722998 NAME:hanws^0.18722998 >> >> >> > > NAME:hehns^0.20172001 NAME:hens^0.2736075 NAME:hins^0.24843 >> >> >> NAME:hons^0.24843 >> >> >> > > NAME:huhns^0.1801875 NAME:huns^0.24843)^2.0) >> >> >> > > > +(((ZIPS:berlin ZIPS:barlin^0.28227 ZIPS:berien^0.25947002 >> >> >> > > ZIPS:berling^0.23232001 ZIPS:perlin^0.26133335))^1.2) >> >> >> > > > >> >> >> > > > The question is just to get some hints where I should look... >> >> >> > > > >> >> >> > > > Both fealds are without norms, omitTf(true) , RAMDirectory, using >> >> >> > > > TopDocs top = ixSearcher.search(q, null, >> >> >> > > > getMaxNumOfCandidates()); >> >> >> > > > and BooleanQuery.setAllowDocsOutOfOrder(true); >> >> >> > > > >> >> >> > > > maybe we made some mistakes on measuring, but we did simple >> >> >> > > > timing >> here >> >> on >> >> >> >> >> >> > > search() method... strange. I would bet it is something we did, >> >> >> > > but I >> >> cannot >> >> >> >> >> >> > see >> >> >> > > where ... >> >> >> > > > >> >> >> > > > >> >> >> > > > >> >> >> > > > >> >> >> > > > >> >> >> > > > >> --------------------------------------------------------------------- >> >> >> > > > To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org >> >> >> > > > For additional commands, e-mail: java-user-h...@lucene.apache.org >> >> >> > > > >> >> >> > > > >> >> >> > > >> >> >> > > --------------------------------------------------------------------- >> >> >> > > To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org >> >> >> > > For additional commands, e-mail: java-user-h...@lucene.apache.org >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> --------------------------------------------------------------------- >> >> >> To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org >> >> >> For additional commands, e-mail: java-user-h...@lucene.apache.org >> >> > >> >> > >> >> > >> >> > >> >> > >> >> > --------------------------------------------------------------------- >> >> > To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org >> >> > For additional commands, e-mail: java-user-h...@lucene.apache.org >> >> > >> >> > >> >> >> >> --------------------------------------------------------------------- >> >> To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org >> >> For additional commands, e-mail: java-user-h...@lucene.apache.org >> > >> > >> > >> > >> > >> > --------------------------------------------------------------------- >> > To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org >> > For additional commands, e-mail: java-user-h...@lucene.apache.org >> > >> > >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org >> For additional commands, e-mail: java-user-h...@lucene.apache.org > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org > For additional commands, e-mail: java-user-h...@lucene.apache.org > > --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
