[ http://issues.apache.org/jira/browse/DERBY-2130?page=all ]
A B updated DERBY-2130:
-----------------------
Attachment: jumpReset.patch
> That is, once I saw 500 seconds for the prepare on database 6, but I
> reran DB 6 and saw 168 seconds. [...] I don't think that the non-
> determinism is "tied" to a particular database and set of tables;
> the same database and tables can sometimes lead to one prepare time,
> and sometimes to a different prepare time.
As I was looking at this problem I noticed a bug in the code that *may* be
causing this variance. It looks like we are not resetting the permuteState
field in OptimizerImpl at the beginning of each "round" of optimization. The
result (in the cases I saw) was that the field has the value WALK_LOW at the
start of the second round of optimization, which, in very timing- sensitive
situations, can lead to a code-path that causes the bestCost for the round to
remain UNSET--i.e. it sits as Double.MAX_VALUE. Then if we try to add anything
to that cost, we end up with Infinity, which effectively ruins most subsequent
cost estimates.
I still have not been able to reproduce the variance that Bryan reports, so I
cannot say for sure that this is the cause. I ran the repro 10 times against
the same database and I consistently saw times of 550 to 700 seconds--and in
each case I could see that the optimizer tried out all 40k permutations before
completing. Then I added code to reset the permuteState variable (see attached
jumpReset.patch) and ran the repro 10 more times against the same database. In
this case every execution completed in about 200 to 220 seconds, and I could
see that the optimizer timed out before trying out all permutations (which is
more what I would expect).
The reason I bring this is up is because the compilation time without
jumpReset.patch and the compilation time with it appear to correlate to the
times that Bryan is seeing in his variance (based on his previous comment). So
I'm wondering if the two are related...?
Bryan, I know you are busy and that running these tests takes a while, but if
you have the time/inclination, can you apply jumpReset.patch to your codeline
and run your tests again? I'm wondering what effect (if any) the patch will
have on the variance that you are seeing...
> Optimizer performance slowdown from 10.1 to 10.2
> ------------------------------------------------
>
> Key: DERBY-2130
> URL: http://issues.apache.org/jira/browse/DERBY-2130
> Project: Derby
> Issue Type: Bug
> Components: Performance, SQL
> Affects Versions: 10.2.1.6, 10.1.3.1, 10.3.0.0
> Reporter: Bryan Pendleton
> Attachments: jumpReset.patch, plan10_1_2_1.txt, plan10_2.txt,
> plans.diff, repro.sql
>
>
> Attached is 'repro.sql', an IJ script which demonstrates what I
> believe to be a serious performance issue in the Optimizer.
> I have run this script in a number of configurations:
> - 10.1.2.1: the script runs successfully. The 'prepare' statement
> takes about 90 seconds, on a fairly powerful Windows machine
> - 10.1.3.1: the script produces a NPE. I believe this is DERBY-1777
> - 10.2.1.8/trunk: the script runs successfully. The 'prepare' statement
> often takes about 220 seconds, on the same Windows machine
> Intermittently, on 10.2 and on the trunk, the prepare statement takes
> 15+ minutes. I cannot reliably reproduce this; I run the same script
> several times in a row and I cannot predict whether it will take 220
> seconds or whether it will take 15+ minutes.
> I am quite motivated to work on this problem, as this is blocking me from
> using Derby for a project that I'm quite keen on, but I need some
> suggestions and ideas about how to attack it. From my perspective
> there are 3 primary topics:
> 1) Why did optimizer performance for this query degrade so significantly
> from 10.1.2.1 to 10.2? The optimizer seems to be at least 2.5 times slower,
> for this particular query at least, in 10.2. Sometimes it is 10x slower.
> 2) What is the source of the non-determinism? Why does the optimizer
> often take 4 minutes to optimize this query on the trunk, but sometimes
> take 15+ minutes? I don't believe that I'm changing anything from
> run to run.
> 3) Can we improve the optimizer performance even beyond what it was
> for 10.1.2? I realize that this is an ugly query, but I was hoping to
> see an optimization time of 5-10 seconds, not 90 seconds (and certainly
> not 220 seconds).
> I have attempted to start answering some of these questions, with
> limited success. Here is some of what I think I've discovered so far:
> - the optimizer changes in 10.2 seem to have given the optimizer many
> more choices of possible query plans to consider. I think this means
> that, if the optimizer does not time out, it will spend substantially
> more time optimizing because there are more choices to evaluate. Does
> this by itself mean that the optimizer will take 2.5 times longer in
> 10.2 than it did in 10.1?
> - something about this query seems to make the costing mechanism go
> haywire, and produce extreme costs. While stepping through the
> optimization of this query in the debugger I have seen it compute
> costs like 1e63 and 1e200. This might be very closely related to
> DERBY-1905, although I don't think I'm doing any subqueries here.
> But maybe I'm misunderstanding the term "subquery" in DERBY-1905.
> At any rate, due to the enormous estimated costs, timeout does not
> occur.
> - the WHERE clause in this query is converted during compilation to
> an equivalent IN clause, I believe, which then causes me to run into
> a number of the problems described in DERBY-47 and DERBY-713.
> Specifically, rather than constructing a plan which involves 4
> index probes for the 4 WHERE clause values, the optimizer decides
> that an index scan must be performed and that it will have to process
> the entire index (because the query uses parameter markers, not
> literal values). So perhaps solving DERBY-47 would help me
> - the optimizer in fact comes up with a "decent" query plan quite quickly.
> I have experimented with placing a hard limit into the optimizer
> timeout code, so that I can force optimization to stop after an
> arbitrary fixed period of time. Then I have been able to set that
> value to as low as 1 second, and the optimizer has produced plans
> that then execute in a few milliseconds. Of course, I have only tried
> this with a trivial amount of data in my database, so it's possible
> that the plan produced by the optimizer after just a second of
> optimizing is in fact poor, and I'm just not noticing it because my
> data sizes are so small.
> At this point, what would be really helpful to me would be some suggestions
> about some general approaches or techniques to try to start breaking down
> and analyzing this problem.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
