ALM-12045 Read Packet Dropped Rate Exceeds the Threshold

Updated on 2024-11-29 GMT+08:00

View PDF

Alarm Description

The system checks the read packet dropped rate every 30 seconds. This alarm is generated when the read packet dropped rate exceeds the threshold for multiple times (the default value is 5).

To change the threshold, choose O&M > Alarm > Thresholds > Name of the desired cluster > Host > Network Reading > Read Packet Dropped Rate.

This alarm is cleared when Trigger Count is 1 and the read packet dropped rate is less than or equal to the threshold. This alarm is cleared when Trigger Count is greater than 1 and the read packet dropped rate is less than or equal to 90% of the threshold.

The alarm detection is disabled by default. If you want to enable this function, check whether this function can be enabled based on Checking System Environments.

Alarm Attributes

Alarm ID	Alarm Severity	Alarm Type	Service Type	Auto Cleared
12045	Critical (default threshold: 5%) Major (default threshold: 0.5%)	Communications	FusionInsight Manager	Yes

Alarm ID

Alarm Severity

Alarm Type

Service Type

Auto Cleared

12045

Critical (default threshold: 5%)

Major (default threshold: 0.5%)

Communications

FusionInsight Manager

Yes

Alarm Parameters

Type	Parameter	Description
Location Information	Source	Specifies the cluster or system for which the alarm is generated.
	ServiceName	Specifies the service for which the alarm is generated.
	RoleName	Specifies the role for which the alarm is generated.
	HostName	Specifies the host for which the alarm is generated.
	NetworkCardName	Specifies the network port for which the alarm is generated.
Additional Information	Trigger Condition	Specifies the threshold for triggering the alarm.

Impact on the System

Latency: When the read packet loss rate of the host network exceeds the threshold, request response is slowed down and services are delayed.
Service failure: When the read packet loss rate of the host network exceeds the threshold, requests cannot be properly responded or times out, which may cause job running failures.

Risk warning: In SUSE kernel 3.0 or later or Red Hat 7.2, the system kernel modifies the mechanism for counting the number of dropped read packets. In this case, this alarm may be generated even if the network is running properly, but services are not affected. You are advised to check the system environment first.

Possible Causes

The NICs are bonded in active/standby mode.
The alarm threshold is improperly configured.
The network quality is poor.

Handling Procedure

On FusionInsight Manager, choose O&M > Alarm > Alarms. On the page that is displayed, click in the row containing the alarm, and view the name of the host for which the alarm is generated and the NIC name.

Check whether the NICs are bonded in active/standby mode.

Log in to the alarm node as user omm and run the ls -l /proc/net/bonding command to check whether the /proc/net/bonding directory exists on the node.
- If yes, the bond mode is configured for the node. Go to 3.
```
# ls -l /proc/net/bonding/
total 0
-r--r--r-- 1 root root 0 Oct 11 17:35 bond0
```
- If no, the bond mode is not configured for the node. Go to 5.
```
# ls -l /proc/net/bonding/
ls: cannot access /proc/net/bonding/: No such file or directory
```

Run the cat /proc/net/bonding/bond0 command to check whether the value of Bonding Mode in the configuration file is fault-tolerance.

NOTE:

In the command, bond0 indicates the name of the bond configuration file. Use the file name obtained in 2.

# cat /proc/net/bonding/bond0 
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)

Bonding Mode: fault-tolerance (active-backup)
Primary Slave: eth1 (primary_reselect always)
Currently Active Slave: eth1
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 1
Slave queue ID: 0

Slave Interface: eth1
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 1
Slave queue ID: 0

If yes, the NICs are bonded in active/standby mode. Go to 4.
If no, go to 5.

Check whether the NIC specified by NetworkCardName in the alarm is the standby NIC.
- If yes, the alarm of the standby NIC cannot be automatically cleared. Manually clear the alarm on the alarm management page. No further action is required.
- If no, go to 5.
  NOTE:
  
  To determine the standby NIC, check the /proc/net/bonding/bond0 configuration file. If the NIC name corresponding to NetworkCardName is Slave Interface but not Currently Active Slave (the current active NIC), the NIC is the standby one.

Check whether the threshold is set properly.

Log in to FusionInsight Manager, choose O&M > Alarm > Thresholds > Name of the desired cluster > Host > Network Reading > Read Packet Dropped Rate, and check whether the alarm threshold is configured properly. The default value is 0.5%. You can adjust the threshold as needed.
- If yes, go to 8.
- If no, go to 6.

Choose O&M > Alarm > Thresholds > Name of the desired cluster > Host > Network Reading > Read Packet Dropped Rate. Click Modify in the Operation column to change the threshold.
After 5 minutes, check whether the alarm is cleared.
- If yes, no further action is required.
- If no, go to 8.

Check whether the network connection is normal.

Contact the network administrator to check whether the network is normal.
- If yes, rectify the fault and go to 9.
- If no, go to 10.
After 5 minutes, check whether the alarm is cleared.
- If yes, no further action is required.
- If no, go to 10.

Collect the fault information.

On FusionInsight Manager of the active cluster, choose O&M. In the navigation pane on the left, choose Log > Download.
Select OMS for Service and click OK.
Expand the Hosts dialog box and select the alarm node and the active OMS node.
Click the edit button in the upper right corner, and set Start Date and End Date for log collection to 30 minutes ahead of and after the alarm generation time respectively. Then, click Download.
Contact O&M engineers and provide the collected logs.