[
https://issues.apache.org/jira/browse/HIVE-23955?focusedWorklogId=467452&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-467452
]
ASF GitHub Bot logged work on HIVE-23955:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 06/Aug/20 17:38
Start Date: 06/Aug/20 17:38
Worklog Time Spent: 10m
Work Description: pkumarsinha commented on a change in pull request #1358:
URL: https://github.com/apache/hive/pull/1358#discussion_r466578320
##########
File path: ql/src/java/org/apache/hadoop/hive/ql/exec/repl/DirCopyTask.java
##########
@@ -86,107 +87,59 @@ private boolean checkIfPathExist(Path sourcePath,
UserGroupInformation proxyUser
return proxyUser.doAs((PrivilegedExceptionAction<Boolean>) () ->
sourcePath.getFileSystem(conf).exists(sourcePath));
}
- private int handleException(Exception e, Path sourcePath, Path targetPath,
- int currentRetry, UserGroupInformation
proxyUser) {
- try {
- LOG.info("Checking if source path " + sourcePath + " is missing for
exception ", e);
- if (!checkIfPathExist(sourcePath, proxyUser)) {
- LOG.info("Source path is missing. Ignoring exception.");
- return 0;
- }
- } catch (Exception ex) {
- LOG.warn("Source path missing check failed. ", ex);
- }
- // retry logic only for i/o exception
- if (!(e instanceof IOException)) {
- LOG.error("Unable to copy {} to {}", sourcePath, targetPath, e);
- setException(e);
- return ErrorMsg.getErrorMsg(e.getMessage()).getErrorCode();
- }
-
- if (currentRetry <= MAX_COPY_RETRY) {
- LOG.warn("Unable to copy {} to {}", sourcePath, targetPath, e);
- } else {
- LOG.error("Unable to copy {} to {} even after retrying for {} time",
sourcePath, targetPath, currentRetry, e);
- setException(e);
- return ErrorMsg.REPL_FILE_SYSTEM_OPERATION_RETRY.getErrorCode();
- }
- int sleepTime = FileUtils.getSleepTime(currentRetry);
- LOG.info("Sleep for " + sleepTime + " milliseconds before retry no " +
(currentRetry));
- try {
- Thread.sleep(sleepTime);
- } catch (InterruptedException timerEx) {
- LOG.info("Sleep interrupted", timerEx.getMessage());
- }
- try {
- if (proxyUser == null) {
- proxyUser = Utils.getUGI();
- }
- FileSystem.closeAllForUGI(proxyUser);
- } catch (Exception ex) {
- LOG.warn("Unable to closeAllForUGI for user " + proxyUser, ex);
- }
- return ErrorMsg.getErrorMsg(e.getMessage()).getErrorCode();
- }
-
@Override
public int execute() {
String distCpDoAsUser =
conf.getVar(HiveConf.ConfVars.HIVE_DISTCP_DOAS_USER);
+ Retryable retryable = Retryable.builder()
+ .withHiveConf(conf)
+ .withRetryOnException(IOException.class).build();
+ try {
+ return retryable.executeCallable(() -> {
+ UserGroupInformation proxyUser = null;
+ Path sourcePath = work.getFullyQualifiedSourcePath();
+ Path targetPath = work.getFullyQualifiedTargetPath();
+ try {
+ if
(conf.getBoolVar(HiveConf.ConfVars.REPL_ADD_RAW_RESERVED_NAMESPACE)) {
+ sourcePath =
reservedRawPath(work.getFullyQualifiedSourcePath().toUri());
+ targetPath =
reservedRawPath(work.getFullyQualifiedTargetPath().toUri());
+ }
+ UserGroupInformation ugi = Utils.getUGI();
+ String currentUser = ugi.getShortUserName();
+ if (distCpDoAsUser != null && !currentUser.equals(distCpDoAsUser)) {
+ proxyUser = UserGroupInformation.createProxyUser(
+ distCpDoAsUser, UserGroupInformation.getLoginUser());
+ }
- Path sourcePath = work.getFullyQualifiedSourcePath();
- Path targetPath = work.getFullyQualifiedTargetPath();
- if (conf.getBoolVar(HiveConf.ConfVars.REPL_ADD_RAW_RESERVED_NAMESPACE)) {
- sourcePath = reservedRawPath(work.getFullyQualifiedSourcePath().toUri());
- targetPath = reservedRawPath(work.getFullyQualifiedTargetPath().toUri());
- }
- int currentRetry = 0;
- int error = 0;
- UserGroupInformation proxyUser = null;
- while (currentRetry <= MAX_COPY_RETRY) {
- try {
- UserGroupInformation ugi = Utils.getUGI();
- String currentUser = ugi.getShortUserName();
- if (distCpDoAsUser != null && !currentUser.equals(distCpDoAsUser)) {
- proxyUser = UserGroupInformation.createProxyUser(
- distCpDoAsUser, UserGroupInformation.getLoginUser());
- }
-
- setTargetPathOwner(targetPath, sourcePath, proxyUser);
-
- // do we create a new conf and only here provide this additional
option so that we get away from
- // differences of data in two location for the same directories ?
- // basically add distcp.options.delete to hiveconf new object ?
- FileUtils.distCp(
- sourcePath.getFileSystem(conf), // source file system
- Collections.singletonList(sourcePath), // list of source paths
- targetPath,
- false,
- proxyUser,
- conf,
- ShimLoader.getHadoopShims());
- return 0;
- } catch (Exception e) {
- currentRetry++;
- error = handleException(e, sourcePath, targetPath, currentRetry,
proxyUser);
- if (error == 0) {
- return 0;
- }
- } finally {
- if (proxyUser != null) {
+ setTargetPathOwner(targetPath, sourcePath, proxyUser);
try {
- FileSystem.closeAllForUGI(proxyUser);
- } catch (IOException e) {
- LOG.error("Unable to closeAllForUGI for user " + proxyUser, e);
- if (error == 0) {
- setException(e);
- error = ErrorMsg.getErrorMsg(e.getMessage()).getErrorCode();
+ if (!checkIfPathExist(sourcePath, proxyUser)) {
+ LOG.info("Source path is missing. Ignoring exception.");
+ return 0;
}
- break;
+ } catch (Exception ex) {
+ LOG.warn("Source path missing check failed. ", ex);
Review comment:
I think we should not. Earlier in this exception case, it wouldn't have
attempted distCp, but now after this change it will.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 467452)
Time Spent: 1.5h (was: 1h 20m)
> Classification of Error Codes in Replication
> --------------------------------------------
>
> Key: HIVE-23955
> URL: https://issues.apache.org/jira/browse/HIVE-23955
> Project: Hive
> Issue Type: Task
> Reporter: Aasha Medhi
> Assignee: Aasha Medhi
> Priority: Major
> Labels: pull-request-available
> Attachments: HIVE-23955.01.patch, HIVE-23955.02.patch, Retry Logic
> for Replication.pdf
>
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
