ALM-45440 Inconsistency Between ClickHouse Replicas

Alarm Description

When the number of ClickHouse replicas is greater than 1, the system periodically checks the replicated table. This alarm is generated if replicated table data is not synchronized. This alarm is cleared when data in all replicated tables between replicas becomes synchronized.

Alarm Attributes

Alarm Parameters

Impact on the System

The data reliability of the ClickHouse replicated table is affected, causing data differences and affecting the query result of the distributed table.

Possible Causes

• The ClickHouse service is overloaded.
• The connection between the ClickHouse and ZooKeeper is abnormal.

Handling Procedure

Check whether the ClickHouse service load is heavy.

1. Log in to FusionInsight Manager, choose O&M > Alarm > Alarms, and view the database name, table name, role name and IP address for the hostname in Location.
2. Log in to the node where the client is installed as the client installation user and run the following commands:

   cd {Client installation path}

   source bigdata_env

   • For a cluster with Kerberos authentication enabled (security mode):

     kinit Component service user

     clickhouse client --host IP address of the ClickHouseServer instance that reports the alarm --port 9440 --secure

   • For a cluster with Kerberos authentication disabled (normal mode):

     clickhouse client --host IP address of the ClickHouseServer instance that reports the alarm --user Username --password --port 9000

3. Run the following statement to check whether data is frequently written to the system table. If yes, wait until the service execution is complete and check whether the alarm is cleared.

   SELECT query_id, user, FQDN(), elapsed, query FROM system.processes ORDER BY query_id;

   • If yes, no further action is required.
   • If no, go to 4.

4. Check whether a large amount of data is written. If yes, wait until the task is complete and check whether the alarm is cleared.
   • If yes, no further action is required.
   • If no, go to 5.

5. Run the following statement to check whether replicas are synchronized:

   select table,absolute_delay, queue_size, inserts_in_queue, merges_in_queue from system.replicas where absolute_delay > 0 order by absolute_delay desc limit 10;

   • If yes, go to 6.
   • If no, go to 9.

6. If inserts_in_queue contains a large amount of content to be inserted, run the following SQL statement to query the replica synchronization queue and locate the error cause:

   SELECT database,table,type,any(last_exception),any(postpone_reason),min(create_time),max(last_attempt_time),max(last_postpone_time),max(num_postponed) AS max_postponed,max(num_tries) AS max_tries,min(num_tries) AS min_tries,countIf(last_exception != '') AS count_err,countIf(num_postponed > 0) AS count_postponed,countIf(is_currently_executing) AS count_executing,count() AS count_all FROM system.replication_queue GROUP BY database,table,type ORDER BY count_all DESC

   Check whether an error message similar to the following is displayed:

   Not executing fetch of part xxx because n fetches already executing, max n

   • If yes, go to 7.
   • If no, go to 9.

7. On FusionInsight Manager, choose Cluster > Services > ClickHouse > Configurations > All Configurations, and check whether the value of background_pool_size is twice the number of cores on the node.
   • If yes, go to 9.
   • If no, go to 8.

8. Set this parameter to twice the number of cores on the node and synchronize the configuration. Wait for a while and check whether the alarm is cleared.
   • If yes, no further action is required.
   • If no, go to 9.

Check the connectivity between ClickHouse and ZooKeeper.

9. Log in to the node where the ClickHouseServer instance is located, go to ${BIGDATA_HOME}/FusionInsight_ClickHouse_*/*_*_ClickHouseServer/etc, and check whether the port numbers of the ClickHouseServer and ZooKeeper in the config.xml file are the same, as shown in the following information in bold:
   

   To view the ZooKeeper port number, choose Cluster > Services > ZooKeeper > Configurations > All Configurations on FusionInsight Manager, and check the value of clientPort.

     <zookeeper>
       <session_timeout_ms>10000</session_timeout_ms>
       <node index="1">
         <host>server-2110082001-0019</host>
         <port>24002</port>
       </node>
       <node index="2">
         <host>server-2110082001-0018</host>
         <port>24002</port>
       </node>
       <node index="3">
         <host>server-2110082001-0017</host>
         <port>24002</port>
       </node>
    </zookeeper>

   • If yes, go to 11.
   • If no, go to 10.

10. Change the port number to the ZooKeeper port number, restart the ClickHouseServer instance, and check whether the alarm is cleared.
    • If yes, no further action is required.
    • If no, go to 11.
      

      During the instance restart, the instance is unavailable, and the ClickHouse service on that instance fails to be executed.

Collect fault information.

11. On FusionInsight Manager, choose O&M. In the navigation pane on the left, choose Log > Download.
12. Expand the Service drop-down list, and select ClickHouse for the target cluster.
13. Expand the Hosts drop-down list. In the Select Host dialog box that is displayed, select the abnormal host, and click OK.
14. Click the edit icon in the upper right corner, and set Start Date and End Date for log collection to 1 hour ahead of and after the alarm generation time, respectively. Then, click Download.
15. Contact O&M personnel and provide the collected logs.

Alarm Clearance

This alarm is automatically cleared after the fault is rectified.

Related Information

None.