[ https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15688165#comment-15688165 ]
stack commented on HBASE-17072: ------------------------------- Ok. Looks like I had my test counts wrong way around. I see that the current master branch does about 1/3rd as many more seeks as the patch does. Here is the seek that current master does that does not happen with the patch (I have seeks throwing a RuntimeException in a LOG). It happens 143 times in a small YCSB run that spans loading and workload c: {code} 26262 2016-11-22 13:40:37,479 INFO [regionserver/ve0528.halxg.cloudera.com/10.17.240.22:16020-longCompactions-1479850170216] hfile.HFileBlock: COUNT=583 26263 java.lang.RuntimeException 26264 at org.apache.hadoop.hbase.io.hfile.HFileBlock$FSReaderImpl.readBlockDataInternal(HFileBlock.java:1710) 26265 at org.apache.hadoop.hbase.io.hfile.HFileBlock$FSReaderImpl.readBlockData(HFileBlock.java:1564) 26266 at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl.readBlock(HFileReaderImpl.java:1484) 26267 at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.readAndUpdateNewBlock(HFileReaderImpl.java:1124) 26268 at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.seekTo(HFileReaderImpl.java:1112) 26269 at org.apache.hadoop.hbase.regionserver.StoreFileScanner.seekAtOrAfter(StoreFileScanner.java:286) 26270 at org.apache.hadoop.hbase.regionserver.StoreFileScanner.seek(StoreFileScanner.java:182) 26271 at org.apache.hadoop.hbase.regionserver.StoreScanner.seekScanners(StoreScanner.java:386) 26272 at org.apache.hadoop.hbase.regionserver.StoreScanner.<init>(StoreScanner.java:294) 26273 at org.apache.hadoop.hbase.regionserver.StoreScanner.<init>(StoreScanner.java:250) 26274 at org.apache.hadoop.hbase.regionserver.compactions.Compactor.createScanner(Compactor.java:515) 26275 at org.apache.hadoop.hbase.regionserver.compactions.Compactor$1.createScanner(Compactor.java:241) 26276 at org.apache.hadoop.hbase.regionserver.compactions.Compactor.compact(Compactor.java:300) 26277 at org.apache.hadoop.hbase.regionserver.compactions.DefaultCompactor.compact(DefaultCompactor.java:66) 26278 at org.apache.hadoop.hbase.regionserver.DefaultStoreEngine$DefaultCompactionContext.compact(DefaultStoreEngine.java:126) 26279 at org.apache.hadoop.hbase.regionserver.HStore.compact(HStore.java:1268) 26280 at org.apache.hadoop.hbase.regionserver.HRegion.compact(HRegion.java:2011) 26281 at org.apache.hadoop.hbase.regionserver.CompactSplitThread$CompactionRunner.doCompaction(CompactSplitThread.java:526) 26282 at org.apache.hadoop.hbase.regionserver.CompactSplitThread$CompactionRunner.run(CompactSplitThread.java:565) 26283 at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) 26284 at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) 26285 at java.lang.Thread.run(Thread.java:745) {code} The seek that happens with and without patch is the below one which is done on initial open of the file post compact or on flush or on open of a region: {code} 613 2016-11-22 13:29:34,730 INFO [StoreFileOpenerThread-info-1] hfile.HFileBlock: COUNT=1 614 java.lang.RuntimeException 615 at org.apache.hadoop.hbase.io.hfile.HFileBlock$FSReaderImpl.readBlockDataInternal(HFileBlock.java:1710) 616 at org.apache.hadoop.hbase.io.hfile.HFileBlock$FSReaderImpl.readBlockData(HFileBlock.java:1564) 617 at org.apache.hadoop.hbase.io.hfile.HFileBlock$FSReaderImpl$2.nextBlock(HFileBlock.java:1464) 618 at org.apache.hadoop.hbase.io.hfile.HFileBlock$FSReaderImpl$2.nextBlockWithBlockType(HFileBlock.java:1472) 619 at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl.<init>(HFileReaderImpl.java:209) 620 at org.apache.hadoop.hbase.io.hfile.HFile.pickReaderVersion(HFile.java:506) 621 at org.apache.hadoop.hbase.io.hfile.HFile.createReader(HFile.java:545) 622 at org.apache.hadoop.hbase.regionserver.StoreFileReader.<init>(StoreFileReader.java:94) 623 at org.apache.hadoop.hbase.regionserver.StoreFileInfo.open(StoreFileInfo.java:270) 624 at org.apache.hadoop.hbase.regionserver.StoreFile.open(StoreFile.java:418) 625 at org.apache.hadoop.hbase.regionserver.StoreFile.createReader(StoreFile.java:530) 626 at org.apache.hadoop.hbase.regionserver.StoreFile.createReader(StoreFile.java:520) 627 at org.apache.hadoop.hbase.regionserver.HStore.createStoreFileAndReader(HStore.java:646) 628 at org.apache.hadoop.hbase.regionserver.HStore.access$000(HStore.java:113) 629 at org.apache.hadoop.hbase.regionserver.HStore$1.call(HStore.java:513) 630 at org.apache.hadoop.hbase.regionserver.HStore$1.call(HStore.java:510) 631 at java.util.concurrent.FutureTask.run(FutureTask.java:266) 632 at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) 633 at java.util.concurrent.FutureTask.run(FutureTask.java:266) 634 at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) 635 at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) 636 at java.lang.Thread.run(Thread.java:745) {code} ... even still... unpatched, we do more of the above than w/ patch. Let me update the patch w/ some better commentary since it is tough digging in on this stuff. > CPU usage starts to climb up to 90-100% when using G1GC > ------------------------------------------------------- > > Key: HBASE-17072 > URL: https://issues.apache.org/jira/browse/HBASE-17072 > Project: HBase > Issue Type: Bug > Components: Performance, regionserver > Affects Versions: 1.0.0, 2.0.0, 1.2.0 > Reporter: Eiichi Sato > Attachments: HBASE-17072.master.001.patch, > disable-block-header-cache.patch, mat-threadlocals.png, mat-threads.png, > metrics.png, slave1.svg, slave2.svg, slave3.svg, slave4.svg > > > h5. Problem > CPU usage of a region server in our CDH 5.4.5 cluster, at some point, starts > to gradually get higher up to nearly 90-100% when using G1GC. We've also run > into this problem on CDH 5.7.3 and CDH 5.8.2. > In our production cluster, it normally takes a few weeks for this to happen > after restarting a RS. We reproduced this on our test cluster and attached > the results. Please note that, to make it easy to reproduce, we did some > "anti-tuning" on a table when running tests. > In metrics.png, soon after we started running some workloads against a test > cluster (CDH 5.8.2) at about 7 p.m. CPU usage of the two RSs started to rise. > Flame Graphs (slave1.svg to slave4.svg) are generated from jstack dumps of > each RS process around 10:30 a.m. the next day. > After investigating heapdumps from another occurrence on a test cluster > running CDH 5.7.3, we found that the ThreadLocalMap contain a lot of > contiguous entries of {{HFileBlock$PrefetchedHeader}} probably due to primary > clustering. This caused more loops in > {{ThreadLocalMap#expungeStaleEntries()}}, consuming a certain amount of CPU > time. What is worse is that the method is called from RPC metrics code, > which means even a small amount of per-RPC time soon adds up to a huge amount > of CPU time. > This is very similar to the issue in HBASE-16616, but we have many > {{HFileBlock$PrefetchedHeader}} not only {{Counter$IndexHolder}} instances. > Here are some OQL counts from Eclipse Memory Analyzer (MAT). This shows a > number of ThreadLocal instances in the ThreadLocalMap of a single handler > thread. > {code} > SELECT * > FROM OBJECTS (SELECT AS RETAINED SET OBJECTS value > FROM OBJECTS 0x4ee380430) obj > WHERE obj.@clazz.@name = > "org.apache.hadoop.hbase.io.hfile.HFileBlock$PrefetchedHeader" > #=> 10980 instances > {code} > {code} > SELECT * > FROM OBJECTS (SELECT AS RETAINED SET OBJECTS value > FROM OBJECTS 0x4ee380430) obj > WHERE obj.@clazz.@name = "org.apache.hadoop.hbase.util.Counter$IndexHolder" > #=> 2052 instances > {code} > Although as described in HBASE-16616 this somewhat seems to be an issue in > G1GC side regarding weakly-reachable objects, we should keep ThreadLocal > usage minimal and avoid creating an indefinite number (in this case, a number > of HFiles) of ThreadLocal instances. > HBASE-16146 removes ThreadLocals from the RPC metrics code. That may solve > the issue (I just saw the patch, never tested it at all), but the > {{HFileBlock$PrefetchedHeader}} are still there in the ThreadLocalMap, which > may cause issues in the future again. > h5. Our Solution > We simply removed the whole {{HFileBlock$PrefetchedHeader}} caching and > fortunately we didn't notice any performance degradation for our production > workloads. > Because the PrefetchedHeader caching uses ThreadLocal and because RPCs are > handled randomly in any of the handlers, small Get or small Scan RPCs do not > benefit from the caching (See HBASE-10676 and HBASE-11402 for the details). > Probably, we need to see how well reads are saved by the caching for large > Scan or Get RPCs and especially for compactions if we really remove the > caching. It's probably better if we can remove ThreadLocals without breaking > the current caching behavior. > FWIW, I'm attaching the patch we applied. It's for CDH 5.4.5. -- This message was sent by Atlassian JIRA (v6.3.4#6332)