accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Adam Fuchs (JIRA)" <>
Subject [jira] [Commented] (ACCUMULO-3067) scan performance degrades after compaction
Date Tue, 19 Aug 2014 22:53:18 GMT


Adam Fuchs commented on ACCUMULO-3067:

The code is pretty simple: it writes a bunch of keys sequentially in a loader program, then
a scan program scans in a loop. The only thing non-standard is that I use a very large batch
size (60,000) to improve scan throughput and reduce the number of reseeks.

Josh is correct that HDFS is nowhere near full (is it time to get rid of that incomprehensible
stat on the monitor page yet?). 

Data cache hits are nearly zero before and after _the event_. The uncompressed data does not
fit in the tserver's block cache, and I'm never revisiting a block before it gets ejected.
The compressed size of the table is about 80MB (0.8B per K/V), so that fits cleaning into
the OS's cache. There is no significant disk I/O before or after _the event_. Index caches
are pegged at about 100% before and after _the event_.

> scan performance degrades after compaction
> ------------------------------------------
>                 Key: ACCUMULO-3067
>                 URL:
>             Project: Accumulo
>          Issue Type: Bug
>          Components: tserver
>         Environment: Macbook Pro 2.6 GHz Intel Core i7, 16GB RAM, SSD, OSX 10.9.4, single
tablet server process, single client process
>            Reporter: Adam Fuchs
>         Attachments: Screen Shot 2014-08-19 at 4.19.37 PM.png, accumulo_query_perf_test.tar.gz
> I've been running some scan performance tests on 1.6.0, and I'm running into an interesting
situation in which query performance starts at a certain level and then degrades by ~15% after
an event. The test follows roughly the following scenario:
>  # Single tabletserver instance
>  # Load 100M small (~10byte) key/values into a tablet and let it finish major compacting
>  # Disable the garbage collector (this makes the time to _the event_ longer)
>  # Restart the tabletserver
>  # Repeatedly scan from the beginning to the end of the table in a loop
>  # Something happens on the tablet server, like one of {idle compaction of metadata table,
forced flush of metadata table, forced compaction of metadata table, forced flush of trace
>  # Observe that scan rates dropped by 15-20%
>  # Observe that restarting the scan will not improve performance back to original level.
Performance only gets better upon restarting the tablet server.
> I've been able to get this not to happen by removing iterators from the iterator tree.
It doesn't seem to matter which iterators, but removing a certain number both improves performance
(significantly) and eliminates the degradation problem. The default iterator tree includes:

>  * SourceSwitchingIterator
>  ** VersioningIterator
>  *** SynchronizedIterator
>  **** VisibilityFilter
>  ***** ColumnQualifierFilter
>  ****** ColumnFamilySkippingIterator
>  ******* DeletingIterator
>  ******** StatsIterator
>  ********* MultiIterator
>  ********** MemoryIterator
>  ********** ProblemReportingIterator
>  *********** HeapIterator
>  ************ RFile.LocalityGroupReader
> We can eliminate the weird condition by narrowing the set of iterators to:
>  * SourceSwitchingIterator
>  ** VisibilityFilter
>  *** ColumnFamilySkippingIterator
>  **** DeletingIterator
>  ***** StatsIterator
>  ****** MultiIterator
>  ******* MemoryIterator
>  ******* ProblemReportingIterator
>  ******** HeapIterator
>  ********* RFile.LocalityGroupReader
> There are other combinations that also perform much better than the default. I haven't
been able to isolate this problem to a single iterator, despite removing each iterator one
at a time.
> Anybody know what might be happening here? Best theory so far: the JVM learns that iterators
can be used in a different way after a compaction, and some JVM optimization like JIT compilation,
branch prediction, or automatic inlining stops happening.

This message was sent by Atlassian JIRA

View raw message