[
https://issues.apache.org/jira/browse/HBASE-22701?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16899680#comment-16899680
]
Ankit Singhal commented on HBASE-22701:
---------------------------------------
bq. The log is saying that trying to use dynamical class loader but code might
not do that based on config being turned OFF or we faced the Exception issue
and disabled it.
Agreed, and also can we let the user know that it's coprocessor classes are not
getting loaded because of an exception while initializing the loader and not
because of the issue with the jar, or can we make the new warning little
verbose(probably with ERROR and saying that coprocessor classes may not be
loaded)?
> Better handle invalid local directory for DynamicClassLoader
> ------------------------------------------------------------
>
> Key: HBASE-22701
> URL: https://issues.apache.org/jira/browse/HBASE-22701
> Project: HBase
> Issue Type: Improvement
> Components: regionserver
> Reporter: Josh Elser
> Assignee: Josh Elser
> Priority: Major
> Fix For: 2.3.0, 2.2.1, 2.1.6
>
> Attachments: HBASE-22701.001.patch
>
>
> If you give HBase an {{hbase.local.dir}} (usually, "{{hbase.tmp.dir}}/local")
> which is not writable to it, you will get some weird errors on the scan path.
> I just saw this (again?) with Phoenix.
> Specifically, the first attempt to reference DynamicClassLoader (via
> ProtobufUtil), will result in an ExceptionInInitializationError because the
> unchecked exception coming out of DynamicClassLoader's constructor interrupts
> the loading of {{DynamicClassLoader.class}}.
> {noformat}
> 2019-07-14 06:25:34,284 ERROR
> [RpcServer.Metadata.Fifo.handler=12,queue=0,port=16020]
> coprocessor.MetaDataEndpointImpl: dropTable failed
> org.apache.hadoop.hbase.DoNotRetryIOException:
> java.lang.ExceptionInInitializerError
> at
> org.apache.hadoop.hbase.client.RpcRetryingCallerImpl.translateException(RpcRetryingCallerImpl.java:221)
> at
> org.apache.hadoop.hbase.client.RpcRetryingCallerImpl.callWithoutRetries(RpcRetryingCallerImpl.java:194)
> at
> org.apache.hadoop.hbase.client.ScannerCallableWithReplicas$RetryingRPC.call(ScannerCallableWithReplicas.java:387)
> at
> org.apache.hadoop.hbase.client.ScannerCallableWithReplicas$RetryingRPC.call(ScannerCallableWithReplicas.java:361)
> at
> org.apache.hadoop.hbase.client.RpcRetryingCallerImpl.callWithRetries(RpcRetryingCallerImpl.java:107)
> at
> org.apache.hadoop.hbase.client.ResultBoundedCompletionService$QueueingFuture.run(ResultBoundedCompletionService.java:80)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> Caused by: java.lang.ExceptionInInitializerError
> at
> org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil.toFilter(ProtobufUtil.java:1598)
> at
> org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil.toScan(ProtobufUtil.java:1152)
> at
> org.apache.hadoop.hbase.regionserver.RSRpcServices.newRegionScanner(RSRpcServices.java:2967)
> at
> org.apache.hadoop.hbase.regionserver.RSRpcServices.scan(RSRpcServices.java:3301)
> at
> org.apache.hadoop.hbase.client.ScannerCallable.openScanner(ScannerCallable.java:332)
> at
> org.apache.hadoop.hbase.client.ScannerCallable.rpcCall(ScannerCallable.java:242)
> at
> org.apache.hadoop.hbase.client.ScannerCallable.rpcCall(ScannerCallable.java:58)
> at
> org.apache.hadoop.hbase.client.RegionServerCallable.call(RegionServerCallable.java:127)
> at
> org.apache.hadoop.hbase.client.RpcRetryingCallerImpl.callWithoutRetries(RpcRetryingCallerImpl.java:192)
> ... 7 more
> Caused by: java.lang.RuntimeException: Failed to create local dir
> /hadoopfs/fs1/hbase/local/jars, DynamicClassLoader failed to init
> at
> org.apache.hadoop.hbase.util.DynamicClassLoader.initTempDir(DynamicClassLoader.java:110)
> at
> org.apache.hadoop.hbase.util.DynamicClassLoader.<init>(DynamicClassLoader.java:98)
> at
> org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil$ClassLoaderHolder.lambda$static$0(ProtobufUtil.java:261)
> at java.security.AccessController.doPrivileged(Native Method)
> at
> org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil$ClassLoaderHolder.<clinit>(ProtobufUtil.java:260)
> ... 16 more
> {noformat}
> Every subsequent call will result in a NoClassDefFoundError, because we
> already tried to load DynamicClassLoader.class once and failed.
> {noformat}
> 2019-07-14 06:25:34,380 ERROR
> [RpcServer.Metadata.Fifo.handler=2,queue=2,port=16020]
> coprocessor.MetaDataEndpointImpl: dropTable failed
> org.apache.hadoop.hbase.DoNotRetryIOException:
> java.lang.NoClassDefFoundError: Could not initialize class
> org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil$ClassLoaderHolder
> at
> org.apache.hadoop.hbase.client.RpcRetryingCallerImpl.translateException(RpcRetryingCallerImpl.java:221)
> at
> org.apache.hadoop.hbase.client.RpcRetryingCallerImpl.callWithoutRetries(RpcRetryingCallerImpl.java:194)
> at
> org.apache.hadoop.hbase.client.ScannerCallableWithReplicas$RetryingRPC.call(ScannerCallableWithReplicas.java:387)
> at
> org.apache.hadoop.hbase.client.ScannerCallableWithReplicas$RetryingRPC.call(ScannerCallableWithReplicas.java:361)
> at
> org.apache.hadoop.hbase.client.RpcRetryingCallerImpl.callWithRetries(RpcRetryingCallerImpl.java:107)
> at
> org.apache.hadoop.hbase.client.ResultBoundedCompletionService$QueueingFuture.run(ResultBoundedCompletionService.java:80)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> Caused by: java.lang.NoClassDefFoundError: Could not initialize class
> org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil$ClassLoaderHolder
> at
> org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil.toFilter(ProtobufUtil.java:1598)
> at
> org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil.toScan(ProtobufUtil.java:1152)
> at
> org.apache.hadoop.hbase.regionserver.RSRpcServices.newRegionScanner(RSRpcServices.java:2967)
> at
> org.apache.hadoop.hbase.regionserver.RSRpcServices.scan(RSRpcServices.java:3301)
> at
> org.apache.hadoop.hbase.client.ScannerCallable.openScanner(ScannerCallable.java:332)
> at
> org.apache.hadoop.hbase.client.ScannerCallable.rpcCall(ScannerCallable.java:242)
> at
> org.apache.hadoop.hbase.client.ScannerCallable.rpcCall(ScannerCallable.java:58)
> at
> org.apache.hadoop.hbase.client.RegionServerCallable.call(RegionServerCallable.java:127)
> at
> org.apache.hadoop.hbase.client.RpcRetryingCallerImpl.callWithoutRetries(RpcRetryingCallerImpl.java:192)
> ... 7 more
> {noformat}
> The client gets an error about this, and would presumably know that something
> is amiss, but an operator wouldn't potentially see this on their own.
> I see two options:
> # We abort the RegionServer when the DynamicClassLoader fails to run
> # We catch the exception and treat the DynamicClassLoader as disabled (same
> action as if you had set {{hbase.use.dynamic.jars=false}}).
> I want to do #1 so that we don't propagate bogus configuration, but it feels
> a bit "harsh" to do that. I think #2 is the right solution with a
> big-fat-warning.
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)
