ALM-14016 DataNode Direct Memory Usage Exceeds the Threshold

Alarm Description

The system checks the direct memory usage of HDFS every 30 seconds. This alarm is generated when the direct memory usage of DataNode instances exceeds the threshold (90% of the maximum memory).

This alarm is automatically cleared when the direct memory usage is less than the threshold.

Alarm Attributes

Alarm Parameters

Impact on the System

If the available direct memory of DataNode instances is insufficient, a memory overflow may occur and the service breaks down.

Possible Causes

The direct memory of DataNode instances is overused or the direct memory is inappropriately allocated.

Handling Procedure

Check the direct memory usage.

1. On the Home page of FusionInsight Manager, choose O&M > Alarm > Alarms. On the page that is displayed, click the drop-down list in the row containing ALM-14016 DataNode Direct Memory Usage Exceeds the Threshold, and view the role name and IP address of the instance for which the alarm is generated in the Location area.
2. On the Home page of FusionInsight Manager, choose Cluster > Services > HDFS. On the page that is displayed, click the Instance tab. In the instance list, select DataNode (IP address of the instance for which this alarm is generated). Click the drop-down list in the upper right corner of the chart, choose Customize > Resource, and select DataNode Memory to check the direct memory usage.
3. Check whether the used direct memory of a DataNode instance reaches 90% (default threshold) of the maximum direct memory allocated to it.
   • If yes, go to 4.
   • If no, go to 8.

4. On the Home page of FusionInsight Manager, choose Cluster > Services > HDFS. On the page that is displayed, click the Configuration tab then the All Configurations sub-tab, and select DataNode > System. Check whether -XX:MaxDirectMemorySize exists in the GC_OPTS parameter.
   • If yes, go to 5.
   • If no, go to 6.

5. Adjust the value of -XX:MaxDirectMemorySize.
   1. In GC_OPTS, check the value of -Xmx and check whether the node memory is sufficient.
      

      You can determine whether the node memory is sufficient based on the actual environment. For example, you can use the following method:

      Use the IP address to log in to the instance for which the alarm is generated as user root and run the free -g command to check the value of Mem in the free column. The value indicates the available memory of the node. In the following example, the available memory of the node is 4 GB.

                    total        used        free      shared  buff/cache   available
      Mem:            112          48           4          10          58          46
      ......

      If the value of Mem is at least that of -Xmx, the node memory is sufficient. If the value of Mem is less than that of -Xmx, the node memory is insufficient.

      • If yes, change the value of -XX:MaxDirectMemorySize to that of -Xmx.
      • If no, increase -XX:MaxDirectMemorySize to a value no larger than that of Mem.
   2. Save the configuration and restart the DataNode instances.

6. Check whether ALM-14008 DataNode Heap Memory Usage Exceeds the Threshold exists.
   • If yes, rectify the fault by referring to ALM-14008 DataNode Heap Memory Usage Exceeds the Threshold.
   • If no, go to 7.

7. Check whether the alarm is cleared.
   • If yes, no further action is required.
   • If no, go to 8.

Collect the fault information.

8. On FusionInsight Manager, choose O&M. In the navigation pane on the left, choose Log > Download.
9. Expand the Service drop-down list, and select DataNode for the target cluster.
10. Click in the upper right corner, and set Start Date and End Date for log collection to 10 minutes ahead of and after the alarm generation time, respectively. Then, click Download.
11. Contact O&M engineers and provide the collected logs.

Alarm Clearance

This alarm is automatically cleared after the fault is rectified.

Related Information

None.