[ https://issues.apache.org/jira/browse/HDFS-7645?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14597138#comment-14597138 ]
Vinayakumar B commented on HDFS-7645: ------------------------------------- bq. This change is incompatible since we expose RollingUpgradeInfo in the NN's JMX (a public API). As discussed above, rather than being null on finalization, it now sets the finalization time. Oh! Thanks [~andrew.wang] for pointing out. That was a miss. bq. Have we thought about other ways of solving this issue? Else we can change the JMX method to still return null on finalization. Since DN side wanted to differentiate between FINALIZED rollingupgrade status and rolledback status, Setting the finalizetime on finalization. bq. Else we can change the JMX method to still return null on finalization. We can do this if this fix is backported to stable branches. Currently its only available in branch-2. If not so critical to change it back, then we can add a release note indicating the change. Note that, {{ClientProtocol#rollingUpgrade(..)}} also changed to return non-null finalized status as well. > Rolling upgrade is restoring blocks from trash multiple times > ------------------------------------------------------------- > > Key: HDFS-7645 > URL: https://issues.apache.org/jira/browse/HDFS-7645 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode > Affects Versions: 2.6.0 > Reporter: Nathan Roberts > Assignee: Keisuke Ogiwara > Fix For: 2.8.0 > > Attachments: HDFS-7645.01.patch, HDFS-7645.02.patch, > HDFS-7645.03.patch, HDFS-7645.04.patch, HDFS-7645.05.patch, > HDFS-7645.06.patch, HDFS-7645.07.patch > > > When performing an HDFS rolling upgrade, the trash directory is getting > restored twice when under normal circumstances it shouldn't need to be > restored at all. iiuc, the only time these blocks should be restored is if we > need to rollback a rolling upgrade. > On a busy cluster, this can cause significant and unnecessary block churn > both on the datanodes, and more importantly in the namenode. > The two times this happens are: > 1) restart of DN onto new software > {code} > private void doTransition(DataNode datanode, StorageDirectory sd, > NamespaceInfo nsInfo, StartupOption startOpt) throws IOException { > if (startOpt == StartupOption.ROLLBACK && sd.getPreviousDir().exists()) { > Preconditions.checkState(!getTrashRootDir(sd).exists(), > sd.getPreviousDir() + " and " + getTrashRootDir(sd) + " should not > " + > " both be present."); > doRollback(sd, nsInfo); // rollback if applicable > } else { > // Restore all the files in the trash. The restored files are retained > // during rolling upgrade rollback. They are deleted during rolling > // upgrade downgrade. > int restored = restoreBlockFilesFromTrash(getTrashRootDir(sd)); > LOG.info("Restored " + restored + " block files from trash."); > } > {code} > 2) When heartbeat response no longer indicates a rollingupgrade is in progress > {code} > /** > * Signal the current rolling upgrade status as indicated by the NN. > * @param inProgress true if a rolling upgrade is in progress > */ > void signalRollingUpgrade(boolean inProgress) throws IOException { > String bpid = getBlockPoolId(); > if (inProgress) { > dn.getFSDataset().enableTrash(bpid); > dn.getFSDataset().setRollingUpgradeMarker(bpid); > } else { > dn.getFSDataset().restoreTrash(bpid); > dn.getFSDataset().clearRollingUpgradeMarker(bpid); > } > } > {code} > HDFS-6800 and HDFS-6981 were modifying this behavior making it not completely > clear whether this is somehow intentional. -- This message was sent by Atlassian JIRA (v6.3.4#6332)