[HDFS-12834] DFSZKFailoverController on error exits with 0 error code

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 2.7.3, 3.0.0-alpha4
• Component/s: ha
• Assignee: Bharat Viswanadham
• Reporter: Zbigniew Kostrzewa
• Created: 17/Nov/17 16:10
• Updated: 05/Dec/17 14:39
• Resolved: 05/Dec/17 14:39

Description

On error DFSZKFailoverController exits with 0 return code which leads to problems when integrating it with scripts and monitoring tools, e.g. systemd, which when configured to restart the service only on failure does not restart ZKFC because it exited with 0.

For example, in my case, systemd reported zkfc exited with success but in logs I have found this:

2017-11-14 05:33:55,075 INFO org.apache.zookeeper.ClientCnxn: Client session timed out, have not heard from server in 3334ms for sessionid 0x15fb794bd240001, closing socket connection and attempting reconnect
2017-11-14 05:33:55,178 INFO org.apache.hadoop.ha.ActiveStandbyElector: Session disconnected. Entering neutral mode...
2017-11-14 05:33:55,564 INFO org.apache.zookeeper.ClientCnxn: Opening socket connection to server 10.9.4.73/10.9.4.73:2182. Will not attempt to authenticate using SASL (unknown error)
2017-11-14 05:33:55,566 INFO org.apache.zookeeper.ClientCnxn: Socket connection established to 10.9.4.73/10.9.4.73:2182, initiating session
2017-11-14 05:33:55,569 INFO org.apache.zookeeper.ClientCnxn: Session establishment complete on server 10.9.4.73/10.9.4.73:2182, sessionid = 0x15fb794bd240001, negotiated timeout = 5000
2017-11-14 05:33:55,570 INFO org.apache.hadoop.ha.ActiveStandbyElector: Session connected.
2017-11-14 05:33:58,230 INFO org.apache.zookeeper.ClientCnxn: Unable to read additional data from server sessionid 0x15fb794bd240001, likely server has closed socket, closing socket connection and attempting reconnect
2017-11-14 05:33:58,335 INFO org.apache.hadoop.ha.ActiveStandbyElector: Session disconnected. Entering neutral mode...
2017-11-14 05:33:58,402 INFO org.apache.zookeeper.ClientCnxn: Opening socket connection to server 10.9.4.138/10.9.4.138:2181. Will not attempt to authenticate using SASL (unknown error)
2017-11-14 05:33:58,403 INFO org.apache.zookeeper.ClientCnxn: Socket connection established to 10.9.4.138/10.9.4.138:2181, initiating session
2017-11-14 05:33:58,406 INFO org.apache.zookeeper.ClientCnxn: Unable to read additional data from server sessionid 0x15fb794bd240001, likely server has closed socket, closing socket connection and attempting reconnect
2017-11-14 05:33:59,218 INFO org.apache.zookeeper.ClientCnxn: Opening socket connection to server 10.9.4.228/10.9.4.228:2183. Will not attempt to authenticate using SASL (unknown error)
2017-11-14 05:33:59,219 INFO org.apache.zookeeper.ClientCnxn: Socket connection established to 10.9.4.228/10.9.4.228:2183, initiating session
2017-11-14 05:33:59,221 INFO org.apache.zookeeper.ClientCnxn: Unable to read additional data from server sessionid 0x15fb794bd240001, likely server has closed socket, closing socket connection and attempting reconnect
2017-11-14 05:34:01,094 INFO org.apache.zookeeper.ClientCnxn: Opening socket connection to server 10.9.4.73/10.9.4.73:2182. Will not attempt to authenticate using SASL (unknown error)
2017-11-14 05:34:01,094 INFO org.apache.zookeeper.ClientCnxn: Client session timed out, have not heard from server in 1773ms for sessionid 0x15fb794bd240001, closing socket connection and attempting reconnect
2017-11-14 05:34:01,196 FATAL org.apache.hadoop.ha.ActiveStandbyElector: Received stat error from Zookeeper. code:CONNECTIONLOSS. Not retrying further znode monitoring connection errors.
2017-11-14 05:34:02,153 INFO org.apache.zookeeper.ZooKeeper: Session: 0x15fb794bd240001 closed
2017-11-14 05:34:02,154 FATAL org.apache.hadoop.ha.ZKFailoverController: Fatal error occurred:Received stat error from Zookeeper. code:CONNECTIONLOSS. Not retrying further znode monitoring connection errors.
2017-11-14 05:34:02,154 INFO org.apache.zookeeper.ClientCnxn: EventThread shut down
2017-11-14 05:34:05,208 INFO org.apache.hadoop.ipc.Server: Stopping server on 8019
2017-11-14 05:34:05,487 INFO org.apache.hadoop.ipc.Server: Stopping IPC Server listener on 8019
2017-11-14 05:34:05,488 INFO org.apache.hadoop.ipc.Server: Stopping IPC Server Responder
2017-11-14 05:34:05,487 INFO org.apache.hadoop.ha.ActiveStandbyElector: Yielding from election
2017-11-14 05:34:05,488 INFO org.apache.hadoop.ha.HealthMonitor: Stopping HealthMonitor thread
2017-11-14 05:34:05,490 FATAL org.apache.hadoop.hdfs.tools.DFSZKFailoverController: Got a fatal error, exiting now
java.lang.RuntimeException: ZK Failover Controller failed: Received stat error from Zookeeper. code:CONNECTIONLOSS. Not retrying further znode monitoring connection errors.
        at org.apache.hadoop.ha.ZKFailoverController.mainLoop(ZKFailoverController.java:369)
        at org.apache.hadoop.ha.ZKFailoverController.doRun(ZKFailoverController.java:238)
        at org.apache.hadoop.ha.ZKFailoverController.access$000(ZKFailoverController.java:61)
        at org.apache.hadoop.ha.ZKFailoverController$1.run(ZKFailoverController.java:172)
        at org.apache.hadoop.ha.ZKFailoverController$1.run(ZKFailoverController.java:168)
        at org.apache.hadoop.security.SecurityUtil.doAsLoginUserOrFatal(SecurityUtil.java:415)
        at org.apache.hadoop.ha.ZKFailoverController.run(ZKFailoverController.java:168)
        at org.apache.hadoop.hdfs.tools.DFSZKFailoverController.main(DFSZKFailoverController.java:181)

The code that seems responsible is in DFSZKFailoverController.java:

  public static void main(String args[])
      throws Exception {
...
    int retCode = 0;
    try {
      retCode = zkfc.run(parser.getRemainingArgs());
    } catch (Throwable t) {
      LOG.fatal("Got a fatal error, exiting now", t); 
    }   
    System.exit(retCode);
  }

Issue Links

• duplicates: HDFS-11751 DFSZKFailoverController daemon exits with wrong status code
• is duplicated by: HDFS-11751 DFSZKFailoverController daemon exits with wrong status code