This is documentation for MapR Version 5.0. You can also refer to MapR documentation for the latest release.

Skip to end of metadata
Go to start of metadata

This section provides reference for the following views in the MapR Control System:

Dashboard - Summary of cluster health, activity, and usage

The Dashboard displays a summary of information about the cluster in six panes.

Panes include:

  • Cluster Heatmap - the alarms and health for each node, by rack
  • Alarms - a summary of alarms for the cluster
  • Cluster Utilization - CPU, Memory, and Disk Space usage
  • Yarn - the number of running and queued applications, number of Node Managers, used memory, total memory, percent of memory used, CPU's used, CPU's total, percent of CPU's used
  • MapReduce - the number of running and queued jobs, running tasks, running map tasks, running reduce tasks, map task capacity, reduce task capacity, map task prefetch capacity, and blacklisted nodes
  • Services - the number of instances of each service
  • Volumes - the number of available, under-replicated, and unavailable volumes

Links in each pane provide shortcuts to more detailed information. The following sections provide information about each pane.

Cluster Heatmap

The Cluster Heatmap pane displays the health of the nodes in the cluster, by rack. Each node appears as a colored square to show its health at a glance.

If you click on the small wrench icon at the upper right of the Cluster Heatmap pane, a key to the color-coded heatmap display slides into view. At the top of the display, you can set the refresh rate for the display (measured in seconds), as well as the number of columns to display (for example, 20 nodes are displayed across two rows for a 10-column display). Click the wrench icon again to slide the display back out of view.

The left drop-down menu at the top of the pane lets you choose which data is displayed. Some of the choices are shown below.

Heatmap legend by category

The heatmap legend changes depending on the criteria you select from the drop-down menu. All the criteria and their corresponding legends are shown here.

Health

  • Healthy - all services up, MapR-FS and all disks OK, and normal heartbeat
  • Upgrading - upgrade in process
  • Degraded - one or more services down, or no heartbeat for over 1 minute
  • Maintenance - routine maintenance in process
  • Critical - Mapr-FS Inactive/Dead/Replicate, or no heartbeat for over 5 minutes
 Click to see the legend for all Heatmap displays, such as CPU, memory and disk space...

CPU Utilization

  • CPU < 50%
  • CPU < 80%
  • CPU >= 80%
  • Unknown

Memory Utilization

  • Memory < 50%
  • Memory < 80%
  • Memory >= 80%
  • Unknown

Disk Space Utilization

  • Used < 50%
  • Used < 80%
  • Used >= 80%
  • Unknown

Too Many Containers Alarm

  • Containers within limit
  • Containers exceeded limit

Duplicate HostId Alarm

  • No duplicate host id detected
  • Duplicate host id detected

UID Mismatch Alarm

  • No UID mismatch detected
  • UID mismatch detected

No Heartbeat Detected Alarm

  • Node heartbeat detected
  • Node heartbeat not detected

TaskTracker Local Dir Full Alarm

  • TaskTracker local directory is not full
  • TaskTracker local directory full

PAM Misconfigured Alarm

  • PAM configured
  • PAM misconfigured

High FileServer Memory Alarm

  • Fileserver memory OK
  • Fileserver memory high

Cores Present Alarm

  • No core files
  • Core files present

Installation Directory Full Alarm

  • Installation Directory free
  • Installation Directory full

Metrics Write Problem Alarm

  • Metrics writing to Database
  • Metrics unable to write to Database

Root Partition Full Alarm

  • Root partition free
  • Root partition full

HostStats Down Alarm

  • HostStats running
  • HostStats down

Webserver Down Alarm

  • Webserver running
  • Webserver down

NFS Gateway Down Alarm

  • NFS Gateway running
  • NFS Gateway down

HBase RegionServer Down Alarm

  • HBase RegionServer running
  • HBase RegionServer down

HBase Master Down Alarm

  • HBase Master running
  • HBase Master down

TaskTracker Down Alarm

  • TaskTracker running
  • TaskTracker down

JobTracker Down Alarm

  • JobTracker running
  • JobTracker down

FileServer Down Alarm

  • FileServer running
  • FileServer down

CLDB Down Alarm

  • CLDB running
  • CLDB down

Time Skew Alarm

  • Time OK
  • Time skew alarm(s)

Software Installation & Upgrades Alarm

  • Version OK
  • Version alarm(s)

Disk Failure(s) Alarm

  • Disks OK
  • Disk alarm(s)

Excessive Logging Alarm

  • No debug
  • Debugging

Zoomed view

You can see a zoomed view of all the nodes in the cluster by moving the zoom slide bar. The zoomed display reveals more details about each node, based on the criteria you chose from the drop-down menu. In this example, CPU Utilization is displayed for each node.

Clicking a rack name navigates to the Nodes view, which provides more detailed information about the nodes in the rack.

Clicking a colored square navigates to the Node Properties View, which provides detailed information about the node.

Alarms

The Alarms pane includes these four columns:

  • Alarm - a list of alarms raised on the cluster
  • Last Raised - the most recent time each alarm state changed
  • Summary - how many nodes or volumes have raised each alarm
  • Clear Alarm - clicking on the X clears the corresponding alarm

Clicking Alarm, Last Raised, or Summary sorts data in ascending or descending order by that column.

Cluster Utilization

The Cluster Utilization pane displays a summary of the total usage of the following resources:

  • CPU
  • Memory
  • Disk Space

For each resource type, the pane displays the percentage of cluster resources used, the amount used, and the total amount present in the system.

A colored dot after the pane's title summarizes the status of the disk and role Configuring Balancer Settings:

  • Green: Both balancers are running.
  • Orange: The replication role balancer is running.
  • Yellow: The disk space balancer is running.
  • Purple: Neither balancer is running.

Click the colored dot to bring up the Balancer Configuration dialog.

Yarn

The Yarn pane shows information about Yarn applications:

  • Running Applications - the number of Yarn applications currently running
  • Queued Applications - the number of Yarn applications queued to run
  • Number of Node Managers - the number of Node Managers (and the number of nodes) in the cluster
  • Used Memory - how much memory has been used to run the applications
  • Total Memory - how much total memory is available for running applications
  • Percent of Memory Used - the percent of memory used compared to the total memory available
  • CPU's Used - the number of CPU cores used
  • CPU's Total - the total number of CPU cores available
  • Percent of CPU's Used - the percent of CPU's used compared to the total number of CPU cores available

Classic MapReduce (v1)

The Classic MapReduce (v1) pane shows information about MapReduce jobs:

  • Running Jobs - the number of MapReduce jobs currently running
  • Queued Jobs - the number of MapReduce jobs queued to run
  • Running Tasks - the number of Map and Reduce tasks currently running
  • Running Map Tasks - the number of Map tasks currently running
  • Running Reduce Tasks - the number of Reduce tasks currently running
  • Map Task Capacity - the number of map slots available across all nodes in the cluster
  • Reduce Task Capacity - the number of reduce slots available across all nodes in the cluster
  • Map Task Prefetch Capacity - the number of map tasks that can be queued to fill map slots once they become available
  • Blacklisted Nodes - the number of nodes that have been eliminated from the MapReduce pool

Services

The Services pane shows information about the services running on the cluster. For each service, the pane displays the following information:

  • Actv - the number of running instances of the service
  • Stby - the number of instances of the service that are configured and standing by to provide failover
  • Stop - the number of instances of the service that have been intentionally stopped
  • Fail - the number of instances of the service that have failed, indicated by a corresponsing Service Down alarm
  • Total - the total number of instances of the service configured on the cluster

Clicking a service navigates to the Services view.

Volumes

The Volumes pane displays the total number of volumes, and the number of volumes that are mounted and unmounted. For each category, the Volumes pane displays the number, percent of the total, and total size.

Clicking Mounted or Unmounted navigates to the Volumes view.

Nodes - Summary of node information

The Nodes view displays the nodes in the cluster, by rack. The Nodes view contains two panes: the Topology pane and the Nodes pane.  The Topology pane shows the racks in the cluster. Selecting a rack displays that rack's nodes in the Nodes pane to the right. Selecting Cluster displays all the nodes in the cluster.

Clicking any column name sorts data in ascending or descending order by that column.

Selecting the checkbox beside one node makes the following buttons available:

  • Properties - navigates to the Node Properties View, which displays detailed information about a single node.
  • Manage Services - displays the Manage Node Services dialog, which lets you start and stop services on the node.
  • Change Topology - displays the Change Node Topology dialog, which lets you change the topology path for a node.

Note: If a node has a No Heartbeat alarm raised, the Forget Node button is also displayed.

When you click on Forget Node, the following Message appears:

When you click on Manage Services, a dialog is displayed where you can stop, start, or restart the services on the node.

When you click on Change Topology, a dialog is displayed where you can choose a different location for the selected node.

Selecting the checkboxes beside multiple nodes changes the text on the buttons to reflect the number of nodes affected:

The dropdown menu at the top left specifies the type of information to display:

  • Overview - general information about each node
  • Services - services running on each node
  • Performance - information about memory, CPU, I/O and RPC performance on each node
  • Disks - information about disk usage, failed disks, and the MapR-FS heartbeat from each node
  • MapReduce - information about the JobTracker heartbeat and TaskTracker slots on each node
  • NFS Nodes - the IP addresses and Virtual IPs assigned to each NFS node
  • Alarm Status - the status of alarms on each node 

Clicking a node's Hostname navigates to the Node Properties View, which provides detailed information about the node.

Selecting the Filter checkbox displays the Filter toolbar, which provides additional data filtering options.

Each time you select a filtering option, the option is displayed in the window below the filter checkbox. You can add more options by clicking on the .

Overview

The Overview displays the following general information about nodes in the cluster:

  • Hlth - each node's health: healthy, degraded, critical, or maintenance
  • Hostname - the hostname of each node
  • Physical IP(s) - the IP address or addresses associated with each node
  • FS HB - time since each node's last heartbeat to the CLDB
  • Physical Topology - the rack path to each node

Services

The Services view displays the following information about nodes in the cluster:

  • Hlth - eact node's health: healthy, degraded, critical, or maintenance
  • Hostname - the hostname of each node
  • Configured Services - a list of the services specified in the config file
  • Running Services - a list of the services running on each node
  • Physical Topology - each node's physical topology

Performance

The Performance view displays the following information about nodes in the cluster, including:

  • Hlth - each node's health: healthy, degraded, critical, or maintenance
  • Hostname - DNS hostname for the nodes in this cluster
  • Memory - percentage of memory used and the total memory
  • % CPU - percentage of CPU usage on the node
  • # CPUs - number of CPUs present on the node
  • Bytes Received - number of bytes received in 1 second, through all network interfaces on the node
 Click to see all Performance metrics...
  • Bytes Sent - number of bytes sent in 1 second, through all network interfaces on the node
  • # RPCs - number of RPC calls
  • RPC In Bytes - number of RPC bytes received by this node every second
  • RPC Out Bytes - number of RPC bytes sent by this node every second
  • # Disk Reads - number of disk read operations on this node every second
  • # Disk Writes - number of disk write operations on this node every second
  • Disk Read Bytes - number of bytes read from all the disks on this node every second
  • Disk Write Bytes - number of bytes written to all the disks on this node every second
  • # Disks - number of disks on this node
  • Gets - 1m - number of data retrievals (gets) executed on this region's primary node in a 1-minute interval
  • Puts - 1m - number of data writes (puts) executed on this region's primary node in a 1-minute interval
  • Scans - 1m - number of data seeks (scans) executed on this region's primary node in a 1-minute interval

Disks

The Disks view displays the following information about nodes in the cluster:

  • Hlth - each node's health: healthy, degraded, or critical
  • Hostname - the hostname of each node
  • # Bad Disks - the number of failed disks on each node
  • Disk Space - the amount of disk used and total disk capacity, in gigabytes

MapReduce

The MapReduce  view displays the following information about nodes in the cluster:

  • Hlth - each node's health: healthy, degraded, or critical
  • Hostname - the hostname of each node
  • TT Map Slots - the number of map slots on each node
  • TT Map Slots Used - the number of map slots in use on each node
  • TT Reduce Slots - the number of reduce slots on each node
  • TT Reduce Slots Used - the number of reduce slots in use on each node

NFS Nodes

The NFS Nodes view displays the following information about nodes in the cluster:

  • Hlth - each node's health: healthy, degraded, or critical
  • Hostname - the hostname of each node
  • Physical IP(s) - the IP address or addresses associated with each node
  • Virtual IP(s) - the virtual IP address or addresses assigned to each node

Alarm Status

The Alarm Status view displays the following information about nodes in the cluster:

  • Hlth - each node's health: healthy, degraded, critical, or maintenance
  • Hostname - DNS hostname for nodes in this cluster
  • Version Alarm - one or more services on the node are running an unexpected version
  • No Heartbeat Alarm - node is not undergoing maintenance, and no heartbeat is detected for over 5 minutes
  • UID Mismatch Alarm - services in the cluster are being run with different user names (UIDs)
  • Duplicate HostId Alarm - two or more nodes in the cluster have the same host id
 Click to see all Alarm Status alerts...
  • Too Many Containers Alarm - number of containers on this node reached the maximum limit
  • Excess Logs Alarm - debug logging is enabled on the node (debug logging generates enormous amounts of data and can fill up disk space)
  • Disk Failure Alarm - a disk has failed on the node
  • Time Skew Alarm - the clock on the node is out of sync with the master CLDB by more than 20 seconds
  • Root Partition Full Alarm - the root partition ("/") on the node is running out of space (99% full)
  • Installation Directory Full Alarm - the partition /opt/mapr on the node is running out of space (95% full)
  • Core Present Alarm - a service on the node has crashed and created a core dump file
  • High FileServer Memory Alarm - memory consumed by fileserver service on the node is high
  • Pam Misconfigured Alarm - the PAM authentication on the node is configured incorrectly
  • TaskTracker Local Directory Full Alarm - the local directory used by the TaskTracker on the specified node(s) is full, and the TaskTracker cannot operate as a result
  • CLDB Alarm - the CLDB service on the node has stopped running
  • FileServer Alarm - the FileServer service on the node has stopped running
  • JobTracker Alarm - the JobTracker service on the node has stopped running
  • TaskTracker Alarm - the TaskTracker service on the node has stopped running
  • HBase Master Alarm - the HBase Master service on the node has stopped running
  • HBase RegionServer Alarm - the HBase RegionServer service on the node has stopped running
  • NFS Gateway Alarm - the NFS service on the node has stopped running
  • WebServer Alarm - the WebServer service on the node has stopped running
  • HostStats Alarm - the HostStats service has stopped running
  • Metrics write problem Alarm - metric data was not written to the database

Node Properties View - Details about a node

The Node Properties view displays detailed information about a single node in seven collapsible panes:

  • Alarms
  • Machine Performance
  • MapR-FS and Available Disks
  • System Disks
  • Manage Node Services
  • MapReduce
  • DB Gets, Puts, Scans

Buttons:

Alarms

The Alarms pane displays a list of alarms that have been raised on the system, and the following information about each alarm:

  • Alarm - the alarm name
  • Last Raised - the most recent time when the alarm was raised
  • Summary - a description of the alarm
  • Clear Alarm - clicking on the X clears the corresponding alarm

Machine Performance

The Machine Performance pane displays the following information about the node's performance and resource usage since it last reported to the CLDB:

  • Memory Used - the amount of memory in use on the node
  • Disk Used - the amount of disk space used on the node
  • CPU - The number of CPUs and the percentage of CPU used on the node
  • Network I/O - the input and output to the node per second
  • RPC I/O - the number of RPC calls on the node and the amount of RPC input and output
  • Disk I/O - the amount of data read to and written from the disk
  • # Operations - the number of disk reads and writes

MapR-FS and Available Disks

The MapR-FS and Available Disks pane displays the disks on the node and information about each disk.

Information headings include:

  • Status - the status of the disk (healthy, failed, or offline)
  • Mount - whether the disk is mounted (indicated by ) or unmounted
  • Device - the device name
  • File System - the file system on the disk
  • Used - the percentage of memory used out of total memory available on the disk
  • Model # - the model number of the disk
  • Serial # - the serial number of the disk
  • Firmware Version - the version of the firmware being used
  • Add to MAPR-FS - clicking the adds the disk to MAPR-FS storage
  • Remove from MAPR-FS - clicking the displays a dialog that asks you to verify that you want to remove the disk

If you confirm by clicking OK, and data on that disk has not been replicated, a warning dialog appears:

For more information on disk status, and the proper procedure for adding, removing, and replacing disks, see the Managing Disks page.

System Disks

The System Disks pane displays information about disks present and mounted on the node:

  • Status - the status of the disk (healthy, failed, or offline)
  • Mount - whether the disk is mounted (indicated by ) or unmounted
  • Device - the device name
  • File System - the file system on the disk
  • Used - the percentage of memory used out of total memory available on the disk
  • Model # - the model number of the disk
  • Serial # - the serial number of the disk
  • Firmware Version - the version of the firmware being used

Manage Node Services

The Manage Node Services pane displays the status of each service on the node.

  • Service - the name of each service
  • State:
    • Configured: the package for the service is installed and the service is configured for all nodes, but it is not enabled for the particular node
    • Not Configured: the package for the service is not installed and/or the service is not configured (configure.sh has not run)
    • Running: the service is installed, has been started by the warden, and is currently executing
    • Stopped: the service is installed and configure.sh has run, but the service is currently not executing
    • StandBy: the service is installed
    • Failed: the service was running, but terminated unexpectedly
  • Log Path - the path to where each service stores its logs
  • Stop/Start:
    • click on to stop the service
    • click on to start the service
  • Restart - click on to restart the service
  • Log Settings - displays the Trace Activity dialog where you can set the level of logging for a service on a particular node. When you select a log level, all the levels listed above it are included in the log. Levels include:
    • ERROR
    • WARN
    • INFO
    • DEBUG
    • TRACE

You can also start and stop services in the the Manage Node Services dialog, by clicking Manage Services in the Nodes view.

MapReduce

The MapReduce pane displays the number of map and reduce slots used, and the total number of map and reduce slots on the node.

DB Gets, Puts, Scans

The DB Gets, Puts, Scans pane displays the number of gets, puts, and scan operations performed during various time intervals.

Node Heatmap

The Node Heatmap view provides a graphical summary of node status across the cluster. This view displays the same information as the Node Heatmap pane on the Dashboard, without the other panes that appear on the dashboard.

Jobs

The Jobs view displays the data collected by the MapR Metrics service. The Jobs view contains two panes: the chart pane and the data grid. The chart pane displays the data corresponding to the selected metric in histogram form. The data grid lists the jobs running on the cluster.

Click on the wrench icon to slide out a menu of information to display. Choices include:

  • Cumulative Job Combine Input Records
  • Cumulative Job Map Input Bytes
  • Cumulative Job Map Input Records
  • Cumulative Job Map Output Bytes
  • Cumulative Job Map Output Records
  • Cumulative Job Reduce Input Records
 Click to see all Job metrics...
  • Cumulative Job Reduce Output Bytes
  • Cumulative Job Reduce Shuffle Bytes
  • Cumulative Physical Memory
  • Current CPU
  • Current Memory
  • Job Average Map Attempt Duration
  • Job Average Reduce Attempt Duration
  • Job Average Task Duration
  • Job Combine Output Records
  • Job Complete Map Task Count
  • Job Complete Reduce Task Count
  • Job Complete Task Count
  • Job Cumulative CPU
  • Job Data-local Map Tasks
  • Job Duration
  • Job End Time
  • Job Error Count
  • Job Failed Map Task Attempt Count
  • Job Failed Map Task Count
  • Job Failed Reduce Task Attempt Count
  • Job Failed Reduce Task Count
  • Job Failed Task Attempt Count
  • Job Failed Task Count
  • Job Id
  • Job Map CPU
  • Job Map Cumulative Memory Bytes
  • Job Map File Bytes Written
  • Job Map GC Time
  • Job Map Input Bytes/Sec
  • Job Map Input Records/Sec
  • Job Map Output Bytes/Sec
  • Job Map Output Records/Sec
  • Job Map Progress
  • Job Map Reserve Slot Wait
  • Job Map Spilled Records
  • Job Map Split Raw Bytes
  • Job Map Task Attempt Count
  • Job Map Task Count
  • Job Map Tasks Duration
  • Job Map Virtual Memory Bytes
  • Job MapR-FS Map Bytes Read
  • Job MapR-FS Map Bytes Written
  • Job MapR-FS Reduce Bytes Read
  • Job MapR-FS Reduce Bytes Written
  • Job MapR-FS Total Bytes Read
  • Job MapR-FS Total Bytes Written
  • Job Maximum Map Attempt Duration
  • Job Maximum Reduce Attempt Duration
  • Job Maximum Task Duration
  • Job Name
  • Job Non-local Map Tasks
  • Job Rack-local Map Tasks
  • Job Reduce CPU
  • Job Reduce Cumulative Memory Bytes
  • Job Reduce File Bytes Written
  • Job Reduce GC Time
  • Job Reduce Input Groups
  • Job Reduce Input Records/Sec
  • Job Reduce Output Records/Sec
  • Job Reduce Progress
  • Job Reduce Reserve Slot Wait
  • Job Reduce Shuffle Bytes/Sec
  • Job Reduce Spilled Records
  • Job Reduce Split Raw Bytes
  • Job Reduce Task Attempt Count
  • Job Reduce Task Count
  • Job Reduce Tasks Duration
  • Job Reduce Virtual Memory Bytes
  • Job Running Map Task Count
  • Job Running Reduce Task Count
  • Job Running Task Count
  • Job Split Raw Bytes
  • Job Start Time
  • Job Submit Time
  • Job Task Attempt Count
  • Job Total File Bytes Written
  • Job Total GC Time
  • Job Total Spilled Records
  • Job Total Task Count
  • Job User
  • Logs
  • Map Tasks Finish Time
  • Map Tasks Start Time
  • Priority
  • Reduce Tasks Finish Time
  • Reduce Tasks Start Time
  • Status
  • Virtual Memory Bytes

Select the Filter checkbox to display the Filter toolbar, which provides additional data filtering options.
The x-axis: drop-down selector lets you change the display scale of the histogram's X axis between a uniform or logarithmic scale. Hover the cursor over a bar in the histogram to display the Filter and Zoom buttons.

Click the Filter button or click the bar to filter the table below the histogram by the data range corresponding to that bar. The selected bar turns yellow. Hover the cursor over the selected bar to display the Clear Filter and Zoom buttons. Click the Clear Filter button to remove the filter from the data range in the table below the histogram.

Double-click a bar or click the Zoom button to zoom in and display a new histogram that displays metrics constrained to the data range represented by the bar. The data range applied to the metrics data set displays above the histogram.

Click the plus or minus buttons in the filter conditions panel to add or remove filter conditions. Uncheck the Filter checkbox above the histogram to clear the entire filter.

Check the box next to a job in the table below the histogram to enable the View Job button. If the job is still running, checking this box also enables the Kill Job button. Clicking Kill Job will display a confirmation dialog to choose whether or not to terminate the job.

Click the View Job button or click the job name in the table below the histogram to open the Job tab for that job.

The Job Pane

From the main Jobs page, select a job from the list below the histogram and click View Job. You can also click directly on the name of the job in the list. The Job Properties pane displays with the Tasks tab selected by default. This pane has three tabs, Tasks, Charts, and Info. If the job is running, the Kill Job button is enabled.

The Tasks Tab

The Tasks tab has two panes. The upper pane displays histograms of metrics for the tasks and task attempts in the selected job. The lower pane displays a table that lists the tasks and primary task attempts in the selected job. Tasks can be in any of the following states:

  • COMPLETE
  • FAILED
  • KILLED
  • PENDING
  • RUNNING

The table of tasks also lists the following information for each task:

  • Task ID. Click the link to display a table with information about the task attempts for this task.
  • Task type:
    • M: Map
    • R: Reduce
    • TC: Task Cleanup
    • JS: Job Setup
    • JC: Job Cleanup
  • Primary task attempt ID. Click the link to display the task attempt pane for this task attempt.
  • Task starting timestamp
  • Task ending timestamp
  • Task duration
  • Host locality
  • Node running the task. Click the link to display the Node Properties pane for this node.

You can select the following task histogram metrics for this job from the drop-down selector:

  • Task Duration
  • Task Attempt Duration
  • Task Attempt Local Bytes Read
  • Task Attempt Local Bytes Written
  • Task Attempt MapR-FS Bytes Read
 Click to see all Task metrics...
  • Task Attempt MapR-FS Bytes Written
  • Task Attempt Garbage Collection Time
  • Task Attempt CPU Time
  • Task Attempt Physical Memory Bytes
  • Task Attempt Virtual Memory Bytes
  • Map Task Attempt Input Records
  • Map Task Attempt Output Records
  • Map Task Attempt Skipped Records
  • Map Task Attempt Input Bytes
  • Map Task Attempt Output Bytes
  • Reduce Task Attempt Input Groups
  • Reduce Task Attempt Shuffle Bytes
  • Reduce Task Attempt Input Records
  • Reduce Task Attempt Output Records
  • Reduce Task Attempt Skipped Records
  • Task Attempt Spilled Records
  • Combined Task Attempt Input Records
  • Combined Task Attempt Output Records

Uncheck the Show Map Tasks box to hide map tasks. Uncheck the Show Reduce Tasks box to hide reduce tasks. Check the Show Setup/Cleanup Tasks box to display job and task setup and cleanup tasks. Histogram filtering and zoom work in the same way as the Jobs pane.

The Charts Tab

Click the Charts tab to display your job's line chart metrics.

Click the Add chart button to add a new line chart. You can use the X and minus buttons at the top-left of each chart to dismiss or hide the chart.

Line charts can display the following metrics for your job:

  • Cumulative CPU used
  • Cumulative physical memory used
  • Number of failed map tasks
  • Number of failed reduce tasks
  • Number of running map tasks
 Click to see all available Chart metrics...
  • Number of running reduce tasks
  • Number of map task attempts
  • Number of failed map task attempts
  • Number of failed reduce task attempts
  • Rate of map record input
  • Rate of map record output
  • Rate of map input bytes
  • Rate of map output bytes
  • Rate of reduce record output
  • Rate of reduce shuffle bytes
  • Average duration of map attempts
  • Average duration of reduce attempts
  • Maximum duration of map attempts
  • Maximum duration of reduce attempts

The Information Tab

The Information tab of the Job Properties pane displays summary information about the job in three collapsible panes:

  • The MapReduce Framework Counters pane displays information about this job's MapReduce activity.
  • The Job Counters pane displays information about the number of this job's map tasks.
  • The File System Counters pane displays information about this job's interactions with the cluster's file system.

The Task Table

The Task table displays a list of the task attempts for the selected task, along with the following information for each task attempt:

  • Status:
    • RUNNING
    • SUCCEEDED
    • FAILED
    • UNASSIGNED
    • KILLED
    • COMMIT PENDING
    • FAILED UNCLEAN
    • KILLED UNCLEAN
  • Task attempt ID. Click the link to display the task attempt pane for this task attempt.
  • Task attempt type:
    • M: Map
    • R: Reduce
    • TC: Task Cleanup
    • JS: Job Setup
    • JC: Job Cleanup
  • Task attempt starting timestamp
  • Task attempt ending timestamp
  • Task attempt shuffle ending timestamp
  • Task attempt sort ending timestamp
  • Task attempt duration
  • Node running the task attempt. Click the link to display the Node Properties pane for this node.
  • A link to the log file for this task attempt
  • Diagnostic information about this task attempt

The Task Attempt Pane

The Task Attempt pane has two tabs, Info and Charts.

The Task Attempt Info Tab

The Info tab displays summary information about this task attempt in three panes:

  • The MapReduce Framework Counters pane displays information about this task attempt's MapReduce activity.
  • The MapReduce Throughput Counters pane displays information about the I/O performance in Bytes/sec and Records/sec.
  • The File System Counters pane displays information about this task attempt's interactions with the cluster's file system.

The Task Attempt Charts Tab

The Task Attempt Charts tab displays line charts for metrics specific to this task attempt. By default, this tab displays charts for these metrics:

  • Cumulative CPU by Time
  • Physical Memory by Time
  • Virtual Memory by Time

Click the Add chart button to add a new line chart. You can use the X and minus buttons at the top-left of each chart to dismiss or hide the chart.

Line charts can display the following metrics for your task:

  • Combine Task Attempt Input Records
  • Combine Task Attempt Output Records
  • Map Task Attempt Input Bytes
  • Map Task Attempt Input Records
  • Map Task Attempt Output Bytes
 Click to see all available Task Attempt metrics...
  • Map Task Attempt Output Records
  • Map Task Attempt Skipped Records
  • Reduce Task Attempt Input Groups
  • Reduce Task Attempt Input Records
  • Reduce Task Attempt Output Records
  • Reduce Task Attempt Shuffle Bytes
  • Reduce Task Attempt Skipped Records
  • Task Attempt CPU Time
  • Task Attempt Local Bytes Read
  • Task Attempt Local Bytes Written
  • Task Attempt MapR-FS Bytes Read
  • Task Attempt MapR-FS Bytes Written
  • Task Attempt Physical Memory Bytes
  • Task Attempt Spilled Records
  • Task Attempt Virtual Memory Bytes
  • No labels