Duo Zhang created HBASE-29214:
---------------------------------
Summary: Typo in AsyncMasterRequestRpcRetryingCaller which makes
us fail to clear the master stub cache
Key: HBASE-29214
URL: https://issues.apache.org/jira/browse/HBASE-29214
Project: HBase
Issue Type: Bug
Components: asyncclient, Client
Reporter: Duo Zhang
{code}
@Override
protected void doCall() {
addListener(conn.getMasterStub(), (stub, error) -> {
if (error != null) {
onError(error, () -> "Get async master stub failed", err -> {
});
return;
}
resetCallTimeout();
addListener(callable.call(controller, stub), (result, error2) -> {
if (error2 != null) {
onError(error2, () -> "Call to master failed",
err -> clearMasterStubCacheOnError(stub, error2)); // <==== this
line
return;
}
future.complete(result);
});
});
}
{code}
It should be err -> clearMasterStubCacheOnError(stub, err), so in
clearMasterStubCacheOnError we can get the unwrapped exception, and then we can
decide whether to clear the master cache correctly.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
