Member since
03-29-2018
4
Posts
0
Kudos Received
1
Solution
My Accepted Solutions
Title | Views | Posted |
---|---|---|
3055 | 03-30-2018 08:27 AM |
01-23-2020
07:35 AM
It's a bug in Oozie. CoordActionCheckXCommand doesn't take care of SUSPENDED state. It only handles SUCCEEDED, FAILED and KILLED. protected Void execute() throws CommandException {
try {
InstrumentUtils.incrJobCounter(getName(), 1, getInstrumentation());
Status slaStatus = null;
CoordinatorAction.Status initialStatus = coordAction.getStatus();
if (workflowJob.getStatus() == WorkflowJob.Status.SUCCEEDED) {
coordAction.setStatus(CoordinatorAction.Status.SUCCEEDED);
// set pending to false as the status is SUCCEEDED coordAction.setPending(0);
slaStatus = Status.SUCCEEDED;
}
else {
if (workflowJob.getStatus() == WorkflowJob.Status.FAILED) {
coordAction.setStatus(CoordinatorAction.Status.FAILED);
slaStatus = Status.FAILED;
// set pending to false as the status is FAILED coordAction.setPending(0);
}
else {
if (workflowJob.getStatus() == WorkflowJob.Status.KILLED) {
coordAction.setStatus(CoordinatorAction.Status.KILLED);
slaStatus = Status.KILLED;
// set pending to false as the status is KILLED coordAction.setPending(0);
}
else {
LOG.warn("Unexpected workflow " + workflowJob.getId() + " STATUS " + workflowJob.getStatus());
coordAction.setLastModifiedTime(new Date());
CoordActionQueryExecutor.getInstance().executeUpdate(
CoordActionQueryExecutor.CoordActionQuery.UPDATE_COORD_ACTION_FOR_MODIFIED_DATE,
coordAction);
return null;
}
}
}
... View more
03-30-2018
08:27 AM
I have resolved this issue! This issue happens because, HDP 2.6.3 consist Atlas old version, and it does't have appropriate timeout parameters. This bug was fix in https://issues.apache.org/jira/browse/ATLAS-690 I bulid Atlas with master branch and run import with new Jars. And added this parameter : atlas.client.readTimeoutMSecs=600000 And all works propertly, without timeout exeptions.
... View more