[jira] [Commented] (HDFS-9081) False-positive ACK slow log in DFSClient
[ https://issues.apache.org/jira/browse/HDFS-9081?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15647839#comment-15647839 ] static-max commented on HDFS-9081: -- I also get this warning in my client logs, but not on any namenode or datanode. So I think this is a rather annoying bug. I get this error with a client application (Apache Flink) with very low throughput (1MB / minute) and all my other Hadoop applications work without any problem. Any work planned to fix this? > False-positive ACK slow log in DFSClient > > > Key: HDFS-9081 > URL: https://issues.apache.org/jira/browse/HDFS-9081 > Project: Hadoop HDFS > Issue Type: Bug >Affects Versions: 2.6.0 >Reporter: He Tianyi >Assignee: He Tianyi >Priority: Minor > > This issue is related with code below: > {noformat} > if (duration > dfsclientSlowLogThresholdMs > && ack.getSeqno() != Packet.HEART_BEAT_SEQNO) { > DFSClient.LOG > .warn("Slow ReadProcessor read fields took " + duration > + "ms (threshold=" + dfsclientSlowLogThresholdMs + "ms); ack: " > + ack + ", targets: " + Arrays.asList(targets)); > } else if (DFSClient.LOG.isDebugEnabled()) { > DFSClient.LOG.debug("DFSClient " + ack); > } > {noformat} > DFSClient prints slow log when awaited after unexpected amount of time > (usually 3 ms). This is a good indicator for network or I/O performance > issue. > However, there is scenario that this slow log is false-positive, i.e. a > reducer, (StageA) iterates over records with identical key, this takes > arbitrary amount of time, but generates no output. (StageB) Then, it output > arbitrary number of records when meet a different key. > If one StageA lasts more than 3 ms (as the example above), there will be > one or more slow log generated, which is not related to any HDFS performance > issue. > In general cases, user should not expect this, as they could be misguided. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-9081) False-positive ACK slow log in DFSClient
[ https://issues.apache.org/jira/browse/HDFS-9081?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14942008#comment-14942008 ] He Tianyi commented on HDFS-9081: - I think code below in {{waitForAckedSeqno}} is already a good indicator of slow acks: {noformat} long duration = Time.monotonicNow() - begin; if (duration > dfsclientSlowLogThresholdMs) { LOG.warn("Slow waitForAckedSeqno took " + duration + "ms (threshold=" + dfsclientSlowLogThresholdMs + "ms)"); } {noformat} While slow log in {{ResponseProcessor}} indicates either slow data producer or slow ack. I'd suggest more informative slow log. > False-positive ACK slow log in DFSClient > > > Key: HDFS-9081 > URL: https://issues.apache.org/jira/browse/HDFS-9081 > Project: Hadoop HDFS > Issue Type: Bug >Affects Versions: 2.6.0 >Reporter: He Tianyi >Assignee: He Tianyi >Priority: Minor > > This issue is related with code below: > {noformat} > if (duration > dfsclientSlowLogThresholdMs > && ack.getSeqno() != Packet.HEART_BEAT_SEQNO) { > DFSClient.LOG > .warn("Slow ReadProcessor read fields took " + duration > + "ms (threshold=" + dfsclientSlowLogThresholdMs + "ms); ack: " > + ack + ", targets: " + Arrays.asList(targets)); > } else if (DFSClient.LOG.isDebugEnabled()) { > DFSClient.LOG.debug("DFSClient " + ack); > } > {noformat} > DFSClient prints slow log when awaited after unexpected amount of time > (usually 3 ms). This is a good indicator for network or I/O performance > issue. > However, there is scenario that this slow log is false-positive, i.e. a > reducer, (StageA) iterates over records with identical key, this takes > arbitrary amount of time, but generates no output. (StageB) Then, it output > arbitrary number of records when meet a different key. > If one StageA lasts more than 3 ms (as the example above), there will be > one or more slow log generated, which is not related to any HDFS performance > issue. > In general cases, user should not expect this, as they could be misguided. -- This message was sent by Atlassian JIRA (v6.3.4#6332)