[ https://issues.apache.org/jira/browse/HBASE-18628?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16133801#comment-16133801 ]
Hadoop QA commented on HBASE-18628: ----------------------------------- | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || | {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m 25s{color} | {color:blue} Docker mode activated. {color} | | {color:green}+1{color} | {color:green} hbaseanti {color} | {color:green} 0m 0s{color} | {color:green} Patch does not have any anti-patterns. {color} | | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 0s{color} | {color:green} The patch does not contain any @author tags. {color} | | {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m 0s{color} | {color:green} The patch appears to include 2 new or modified test files. {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 4m 28s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 0m 47s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m 50s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} mvneclipse {color} | {color:green} 0m 16s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 3m 27s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 0m 30s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 0m 47s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 0m 41s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 0m 41s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m 48s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} mvneclipse {color} | {color:green} 0m 15s{color} | {color:green} the patch passed {color} | | {color:red}-1{color} | {color:red} whitespace {color} | {color:red} 0m 0s{color} | {color:red} The patch has 1 line(s) that end in whitespace. Use git apply --whitespace=fix <<patch_file>>. Refer https://git-scm.com/docs/git-apply {color} | | {color:green}+1{color} | {color:green} hadoopcheck {color} | {color:green} 34m 47s{color} | {color:green} Patch does not cause any errors with Hadoop 2.6.1 2.6.2 2.6.3 2.6.4 2.6.5 2.7.1 2.7.2 2.7.3 or 3.0.0-alpha4. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 4m 4s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 0m 38s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} unit {color} | {color:green}156m 14s{color} | {color:green} hbase-server in the patch passed. {color} | | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 18s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black}209m 39s{color} | {color:black} {color} | \\ \\ || Subsystem || Report/Notes || | Docker | Client=17.05.0-ce Server=17.05.0-ce Image:yetus/hbase:bdc94b1 | | JIRA Issue | HBASE-18628 | | JIRA Patch URL | https://issues.apache.org/jira/secure/attachment/12882658/HBASE-18628.v2.patch | | Optional Tests | asflicense javac javadoc unit findbugs hadoopcheck hbaseanti checkstyle compile | | uname | Linux ab708dbe8621 3.13.0-116-generic #163-Ubuntu SMP Fri Mar 31 14:13:22 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux | | Build tool | maven | | Personality | /home/jenkins/jenkins-slave/workspace/PreCommit-HBASE-Build@2/component/dev-support/hbase-personality.sh | | git revision | master / 56f9e1a | | Default Java | 1.8.0_144 | | findbugs | v3.1.0-RC3 | | whitespace | https://builds.apache.org/job/PreCommit-HBASE-Build/8162/artifact/patchprocess/whitespace-eol.txt | | Test Results | https://builds.apache.org/job/PreCommit-HBASE-Build/8162/testReport/ | | modules | C: hbase-server U: hbase-server | | Console output | https://builds.apache.org/job/PreCommit-HBASE-Build/8162/console | | Powered by | Apache Yetus 0.4.0 http://yetus.apache.org | This message was automatically generated. > ZKPermissionWatcher blocks all ZK notifications > ----------------------------------------------- > > Key: HBASE-18628 > URL: https://issues.apache.org/jira/browse/HBASE-18628 > Project: HBase > Issue Type: Bug > Components: regionserver > Reporter: Mike Drob > Assignee: Mike Drob > Priority: Critical > Attachments: HBASE-18628.patch, HBASE-18628.v2.patch, jstack > > > Buckle up folks, we're going for a ride here. I've seeing this on a branch-2 > based build, but I think the problem will affect branch-1 as well. I'm not > able to easily reproduce the issue, but it will usually come up within an > hour on a given cluster that I have, at which point the problem persists > until an RS restart. I've been seeing the problem and paying attention for > maybe two months, but I suspect it's been happening much longer than that. > h3. Problem > When running in a secure cluster, sometimes the ZK EventThread will get stuck > on a permissions update and not be able to process new notifications. This > happens to also block flush and snapshot, which is how we found it. > h3. Analysis > The main smoking gun is seeing this in repeated jstacks: > {noformat} > "main-EventThread" #43 daemon prio=5 os_prio=0 tid=0x00007f0b92644000 > nid=0x6e69 waiting on condition [0x00007f0b6730f000] > java.lang.Thread.State: TIMED_WAITING (sleeping) > at java.lang.Thread.sleep(Native Method) > at > org.apache.hadoop.hbase.security.access.ZKPermissionWatcher.nodeChildrenChanged(ZKPermissionWatcher.java:191) > at > org.apache.hadoop.hbase.zookeeper.ZooKeeperWatcher.process(ZooKeeperWatcher.java:503) > at > org.apache.zookeeper.ClientCnxn$EventThread.processEvent(ClientCnxn.java:522) > at > org.apache.zookeeper.ClientCnxn$EventThread.run(ClientCnxn.java:498) > {noformat} > That sleep is a 20ms sleep in an {{AtomicReference.compareAndSet}} loop - but > it never gets past the condition. > {code} > while (!nodes.compareAndSet(null, nodeList)) { > try { > Thread.sleep(20); > } catch (InterruptedException e) { > LOG.warn("Interrupted while setting node list", e); > Thread.currentThread().interrupt(); > } > } > {code} > The warning never shows up in the logs, it just keeps looping and looping. > The last relevant line from the watcher in logs is: > {noformat} > 2017-08-17 21:25:12,379 DEBUG > org.apache.hadoop.hbase.zookeeper.ZooKeeperWatcher: > regionserver:22101-0x15df38884c80024, quorum=zk1:2181,zk2:2181,zk3:2181, > baseZNode=/hbase Received ZooKeeper Event, type=NodeChildrenChanged, > state=SyncConnected, path=/hbase/acl > {noformat} > Which makes sense, because the code snippet is from permission watcher's > {{nodeChildrenChanged}} handler. > The separate thread introduced in HBASE-14370 is present, but not doing > anything. And this event hasn't gotten to the part where it splits off into a > thread: > {noformat} > "zk-permission-watcher4-thread-1" #160 daemon prio=5 os_prio=0 > tid=0x0000000001750800 nid=0x6fd9 waiting on condition [0x00007f0b5dce5000] > java.lang.Thread.State: WAITING (parking) > at sun.misc.Unsafe.park(Native Method) > - parking to wait for <0x00000007436ecea0> (a > java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject) > at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175) > at > java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039) > at > java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442) > at > java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1074) > at > java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1134) > at > java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) > at java.lang.Thread.run(Thread.java:748) > {noformat} > h3. Solutions > There's a few approaches we can take to fix this, I think they are all > complimentary. It might be useful to file subtasks or new issues for some of > the solutions if they are longer term. > # Move flush and snapshot to ProcedureV2. This makes my proximate problem go > away, but it's only relevant to branch-2 and master, and doesn't fix anything > on branch-1. Also, Permissions updates would still get stuck, preventing > future permissions updates. I think this is important long term for the > robustness of the system, but not a viable short term fix. > # Add an Executor to ZookeeperWatcher and launch threads from there. Maybe > we'd want to pull the Executor out of ZKPW, but that's not strictly necessary > and can be optimized later -- if we're already threading, then adding another > layer isn't a huge cost. > # Figure out the race condition or logic problem that causes {{nodes}} to be > non-null above. I've tried looking at this and visual inspection isn't > getting me anywhere. -- This message was sent by Atlassian JIRA (v6.4.14#64029)