HDFS-17232. RBF: Fix NoNamenodesAvailableException for a long time, when use observer. #6208
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
💔 -1 overall
This message was automatically generated. |
KeeProMise
changed the title
|
🎊 +1 overall
This message was automatically generated. |
goiri
changed the title
|
🎊 +1 overall
This message was automatically generated. |
simbadzina
reviewed
Oct 24, 2023
| nsId, listObserversFirst, firstNamenodeId, namenode.getNamenodeId()); | ||
| return rotatedNnContexts; | ||
|
|
||
| // If the last namenode in the cache at this time is the namenode. |
There was a problem hiding this comment.
is the namenode.
Should this be expanded to is the namenode whose priority needs to be lowered?
| } | ||
| return namenodeContexts; | ||
|
|
||
| // Move the abnormal namenode to the end of the cache, |
There was a problem hiding this comment.
thanks for your advice, done
| return namenodeContexts; | ||
|
|
||
| // Move the abnormal namenode to the end of the cache, | ||
| // to ensure that the current namenode will not be accessed first next time. |
There was a problem hiding this comment.
Is the current namenode here the same as the abnormal one in the line above?
There was a problem hiding this comment.
yes, I have modified the comment
...-hdfs-rbf/src/main/java/org/apache/hadoop/hdfs/server/federation/router/RouterRpcClient.java
Show resolved
Hide resolved
| } | ||
|
|
||
| /** | ||
| * Shutdown oberver namenode in the cluste. |
| /** | ||
| * Shutdown oberver namenode in the cluste. | ||
| * | ||
| * @param num The number of shutdown oberver. |
Thanks for reviewing this pr! |
|
🎊 +1 overall
This message was automatically generated. |
goiri
approved these changes
Oct 25, 2023
| @@ -772,13 +779,22 @@ public static boolean isUnavailableException(IOException ioe) { | |||
| * Check if the cluster of given nameservice id is available. | |||
| * | |||
| * @param nsId nameservice ID. | |||
| * @param namenode | |||
| if (nameNode != null && nameNode.isObserverState()) { | ||
| cluster.getCluster().shutdownNameNode(nnIndex); | ||
| num--; | ||
| if (num == 0) { |
There was a problem hiding this comment.
Can you just do it in the for loop?
int numNns = cluster.getNamenodes().size();
for (nnIndex = 0; nnIndex < numNns && num > 0; nnIndex++) {
It is also a little weird to modify an input parameter.
There was a problem hiding this comment.
Thanks for your advice, done
|
🎊 +1 overall
This message was automatically generated. |
goiri
approved these changes
Oct 26, 2023
simbadzina
approved these changes
Nov 15, 2023
Comment on lines
791
to
793
| // Use observer and the namenode that causes the exception is an observer, | ||
| // false is returned so that the observer can be marked as unavailable,so other observers | ||
| // or active namenode which is standby in the cache of the router can be retried. |
There was a problem hiding this comment.
This comment is a bit confusing to read. Maybe the following could make it clearer
If the operation is an observer read and the namenode that caused the exception is an observer ...
There was a problem hiding this comment.
Thanks for your advice.
simbadzina
reviewed
Nov 15, 2023
Comment on lines
256
to
267
| public FileSystem getFileSystemWithConfiguredFailoverProxyProvider() throws IOException { | ||
| conf.set(DFS_NAMESERVICES, | ||
| conf.get(DFS_NAMESERVICES)+ ",router-service"); | ||
| conf.set(DFS_HA_NAMENODES_KEY_PREFIX + ".router-service", "router1"); | ||
| conf.set(DFS_NAMENODE_RPC_ADDRESS_KEY+ ".router-service.router1", | ||
| getFileSystemURI().toString()); | ||
| conf.set(HdfsClientConfigKeys.Failover.PROXY_PROVIDER_KEY_PREFIX | ||
| + "." + "router-service", ConfiguredFailoverProxyProvider.class.getName()); | ||
| DistributedFileSystem.setDefaultUri(conf, "hdfs://router-service"); | ||
|
|
||
| return DistributedFileSystem.get(conf); | ||
| } |
There was a problem hiding this comment.
getFileSystemWithConfiguredFailoverProxyProvider and getFileSystemWithObserverReadProxyProvider() share the same configuration, beside the proxy provider. You can separate shared config these one into a separate function, then just set the proxy provider.
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
@goiri @simbadzina If no more comments here, please help merge it, thanks! |
|
Merged the change. Thanks for the contribution @KeeProMise |
jiajunmao
pushed a commit
to jiajunmao/hadoop-MLEC
that referenced
this pull request
Feb 6, 2024
…hen use observer. (apache#6208)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description of PR
see https://issues.apache.org/jira/browse/HDFS-17232
How was this patch tested?
Added new unit test class TestNoNamenodesAvailableLongTime.
For code changes:
LICENSE,LICENSE-binary,NOTICE-binaryfiles?