### What changes were proposed in this pull request?
Cleans up the pooled send buffers and push tasks if the SendBufferPool has been idle for more than
`celeborn.client.push.sendbufferpool.expireTimeout`.
### Why are the changes needed?
Before this PR the SendBufferPool will cache the send buffers and push tasks forever. If they are large
and will not be reused in the future, it wastes memory and causes GC.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Passes GA and manual tests.
Closes#1735 from waitinfuture/812-1.
Authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Fix some typos and grammar
### Why are the changes needed?
Ditto
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
manually test
Closes#1733 from onebox-li/fix-typo.
Authored-by: onebox-li <lyh-36@163.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
As title.
### Why are the changes needed?
Timeout of ```RpcEndpointRef.ask``` is controlled by ```celeborn.rpc.askTimeout```,
so we also need to increase ```celeborn.rpc.askTimeout``` to extend the timeout of commit files.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Passes GA and manual test.
Closes#1725 from waitinfuture/803-fu.
Authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
As title.
### Why are the changes needed?
In 0.2.1-incubating, commit files default timeout is ```NETWORK_TIMEOUT```, which is 240s.
It's more reasonable because commit files costs relatively long time. In my testing with tough disks,
30s timeout with 2 retires is not enough.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Passes GA and manual test.
Closes#1724 from waitinfuture/803.
Authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
…up client
### What changes were proposed in this pull request?
Add heartbeat from client to lifecycle manager. In this PR heartbeat request contains local shuffle ids from
client, lifecycle manager checks with it's local set and returns ids it doesn't know. Upon receiving response,
client calls ```unregisterShuffle``` for cleanup.
### Why are the changes needed?
Before this PR, client side ```unregisterShuffle``` is never called. When running TPCDS 3T with spark thriftserver
without DRA, I found the Executor's heap contains 1.6 million PartitionLocation objects (and StorageInfo):
After this PR, the number of PartitionLocation objects decreases to 275 thousands
This heartbeat can be extended in the future for other purposes, i.e. reporting client's metrics.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Passes GA and manual test.
Closes#1719 from waitinfuture/798.
Authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
…Flight.total```
### What changes were proposed in this pull request?
Refer to https://github.com/apache/incubator-celeborn/pull/1720#discussion_r1265092164
### Why are the changes needed?
ditto
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Passes GA.
Closes#1723 from waitinfuture/799-fu.
Authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
As title.
### Why are the changes needed?
In case where worker instances is very large, say 1000, then before this PR total memory consumed
by inflight requests is 64K * 1000 * ```celeborn.client.push.maxReqsInFlight(16)``` = 1G. This PR
limits total inflight push requests, as 0.2.1-incubating does.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Passes GA and manual test.
Closes#1720 from waitinfuture/799.
Lead-authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Co-authored-by: Cheng Pan <pan3793@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Now slots is not a bottleneck, change SPARK_SHUFFLE_FORCE_FALLBACK_PARTITION_THRESHOLD default value to Int.MaxValue.
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Closes#1695 from AngersZhuuuu/CELEBORN-780.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
…Flight to 16
### What changes were proposed in this pull request?
Change default value of celeborn.client.push.maxReqsInFlight to 16.
### Why are the changes needed?
Previous value 4 is too small, 16 is more reasonable.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Pass GA.
Closes#1683 from waitinfuture/769.
Authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Make Celeborn leader clean expired app dirs on HDFS when an application is Lost.
### Why are the changes needed?
If Celeborn is working on HDFS, the storage manager starts and cleans expired app directories, and the newly created worker will want to delete any unknown app directories.
This will cause using app directories to be deleted unexpectedly.
### Does this PR introduce _any_ user-facing change?
NO.
### How was this patch tested?
UT and cluster.
Closes#1678 from FMX/CELEBORN-764.
Lead-authored-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
Co-authored-by: Cheng Pan <pan3793@gmail.com>
Co-authored-by: Ethan Feng <fengmingxiao.fmx@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
…memory allocator
### What changes were proposed in this pull request?
Changes the following configs' default values
| config | previous value | current value |
| ------------- | ------------- | ------------- |
| celeborn.network.memory.allocator.share | false | true |
| celeborn.client.shuffle.batchHandleChangePartition.enabled | false | true |
| celeborn.client.shuffle.batchHandleCommitPartition.enabled | false | true |
### Why are the changes needed?
In my test, when graceful shutdown is enabled but ```celeborn.client.shuffle.batchHandleChangePartition.enabled``` and ```celeborn.client.shuffle.batchHandleCommitPartition.enabled``` disabled, the worker takes much longer to stop than the two configs enabled.
In another test where worker size is quite small(2 cores 4 G) and replication is on, if shared allocator is disabled, the netty's onTrim fails to release memory, and further causes push data timeout.
### Does this PR introduce _any_ user-facing change?
No, these conifgs are introduces from 0.3.0.
### How was this patch tested?
Passes GA.
Closes#1682 from waitinfuture/768.
Authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
As title
### Why are the changes needed?
To clarify the usage of conf `celeborn.client.spark.push.sort.memory.threshold`
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Pass GA
Closes#1680 from cfmcgrady/docs.
Authored-by: Fu Chen <cfmcgrady@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Support to decide whether to compress shuffle data through configuration.
### Why are the changes needed?
Currently, Celeborn compresses all shuffle data, but for example, the shuffle data of Gluten has already been compressed. In this case, no additional compression is required. Therefore, configuration needs to be provided for users to decide whether to use Celeborn’s compression according to the actual situation.
### Does this PR introduce _any_ user-facing change?
no.
Closes#1669 from kerwin-zk/celeborn-755.
Authored-by: xiyu.zk <xiyu.zk@alibaba-inc.com>
Signed-off-by: Cheng Pan <chengpan@apache.org>
### What changes were proposed in this pull request?
As title
### Why are the changes needed?
In order to distinguish it from the existing master/worker, refactor data replication terminology to 'primary/replica' for improved clarity and inclusivity in the codebase
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
existing tests
Closes#1639 from cfmcgrady/primary-replica.
Lead-authored-by: Fu Chen <cfmcgrady@gmail.com>
Co-authored-by: Cheng Pan <chengpan@apache.org>
Signed-off-by: Cheng Pan <chengpan@apache.org>
### What changes were proposed in this pull request?
This pull PR is an integral component of #1639 . It primarily focuses on updating configuration settings and metrics terminology, while ensuring compatibility with older client versions by refraining from introducing changes related to RPC.
### Why are the changes needed?
In order to distinguish it from the existing master/worker, refactor data replication terminology to 'primary/replica' for improved clarity and inclusivity in the codebase
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
existing tests.
Closes#1650 from cfmcgrady/primary-replica-metrics.
Authored-by: Fu Chen <cfmcgrady@gmail.com>
Signed-off-by: Cheng Pan <chengpan@apache.org>
### What changes were proposed in this pull request?
Refine the congestion relevant code/log/comments
### Why are the changes needed?
ditto
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
manually test
Closes#1637 from onebox-li/improve-congestion.
Authored-by: onebox-li <lyh-36@163.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Unify exclude and blacklist related configuration
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Closes#1633 from AngersZhuuuu/CELEBORN-666-NEW.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
### What changes were proposed in this pull request?
This PR batches revive requests and periodically send to LifecycleManager to reduce number or RPC requests.
To be more detailed. This PR changes Revive message to support multiple unique partitions, and also passes a set unique mapIds for checking MapEnd. Each time ShuffleClientImpl wants to revive, it adds a ReviveRquest to ReviveManager and wait for result. ReviveManager batches revive requests and periodically send to LifecycleManager (deduplicated by partitionId). LifecycleManager constructs ChangeLocationsCallContext and after all locations are notified, it replies to ShuffleClientImpl.
### Why are the changes needed?
In my test 3T TPCDS q23a with 3 Celeborn workers, when kill a worker, the LifecycleManger will receive 4.8w Revive requests:
```
[emr-usermaster-1-1 logs]$ cat spark-emr-user-org.apache.spark.sql.hive.thriftserver.HiveThriftServer2-1-master-1-1.c-fa08904e94c028d1.out.1 |grep -i revive |wc -l
64364
```
After this PR, number of ReviveBatch requests reduces to 708:
```
[emr-usermaster-1-1 logs]$ cat spark-emr-user-org.apache.spark.sql.hive.thriftserver.HiveThriftServer2-1-master-1-1.c-fa08904e94c028d1.out |grep -i revive |wc -l
2573
```
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manual test. I have tested:
1. Disable graceful shutdown, kill one worker, job succeeds
2. Disable graceful shutdown, kill two workers successively, job fails as expected
3. Enable graceful shutdown, restart two workers successively, job succeeds
4. Enable graceful shutdown, restart two workers successively, then kill the third one, job succeeds
Closes#1588 from waitinfuture/656-2.
Lead-authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Co-authored-by: Keyong Zhou <zhouky@apache.org>
Co-authored-by: Keyong Zhou <waitinfuture@gmail.com>
Signed-off-by: Shuang <lvshuang.tb@gmail.com>
### What changes were proposed in this pull request?
1. Celeborn supports storage type selection. HDD, SSD, and HDFS are available for now.
2. Add new buffer size for HDFS file writers.
3. Worker support empty working dirs.
### Why are the changes needed?
Support HDFS only scenario.
### Does this PR introduce _any_ user-facing change?
NO.
### How was this patch tested?
UT and cluster.
Closes#1619 from FMX/CELEBORN-568.
Lead-authored-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
Co-authored-by: Ethan Feng <fengmingxiao.fmx@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
30s for fetch timeout is too short and easy to exceed. This PR increases the default value to 600s.
### Why are the changes needed?
When I was testing 3T TPCDS with three workers, I encountered fetch timeout:
```
23/06/21 16:46:41,771 INFO [fetch-server-11-7] FetchHandler: Sending chunk 28856864163, 1, 0, 2147483647
...
23/06/21 16:47:16,870 INFO [fetch-server-11-7] FetchHandler: Sent chunk 28856864163, 1, 0, 2147483647
```
And I remember from some users' monitoring, the max fetch time can reach several minutes on heavy load without error.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manual test.
Closes#1618 from waitinfuture/709.
Authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Refresh celeborn configurations in doc
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Closes#1592 from AngersZhuuuu/CELEBORN-680.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
### What changes were proposed in this pull request?
Celeborn fetch chunk also should support check timeout
#### Test case
```
executor instance 20
SQL:
SELECT count(1) from (select /*+ REPARTITION(100) */ * from spark_auxiliary.t50g) tmp;
--conf spark.celeborn.client.spark.shuffle.writer=sort \
--conf spark.celeborn.client.fetch.excludeWorkerOnFailure.enabled=true \
--conf spark.celeborn.client.push.timeout=10s \
--conf spark.celeborn.client.push.replicate.enabled=true \
--conf spark.celeborn.client.push.revive.maxRetries=10 \
--conf spark.celeborn.client.reserveSlots.maxRetries=10 \
--conf spark.celeborn.client.registerShuffle.maxRetries=3 \
--conf spark.celeborn.client.push.blacklist.enabled=true \
--conf spark.celeborn.client.blacklistSlave.enabled=true \
--conf spark.celeborn.client.fetch.timeout=30s \
--conf spark.celeborn.client.push.data.timeout=30s \
--conf spark.celeborn.client.push.limit.inFlight.timeout=600s \
--conf spark.celeborn.client.push.maxReqsInFlight=32 \
--conf spark.celeborn.client.shuffle.compression.codec=ZSTD \
--conf spark.celeborn.rpc.askTimeout=30s \
--conf spark.celeborn.client.rpc.reserveSlots.askTimeout=30s \
--conf spark.celeborn.client.shuffle.batchHandleChangePartition.enabled=true \
--conf spark.celeborn.client.shuffle.batchHandleCommitPartition.enabled=true \
--conf spark.celeborn.client.shuffle.batchHandleReleasePartition.enabled=true
```
Test with 3 worker and add a `Thread.sleep(100s)` before worker handle `ChunkFetchRequest`
Before patch
<img width="1783" alt="截屏2023-06-14 上午11 20 55" src="https://github.com/apache/incubator-celeborn/assets/46485123/182dff7d-a057-4077-8368-d1552104d206">
After patch
<img width="1792" alt="image" src="https://github.com/apache/incubator-celeborn/assets/46485123/3c8b7933-8ace-426d-8e9f-04e0aabfac8e">
The log shows the fetch timeout checker workers
```
23/06/14 11:14:54 ERROR WorkerPartitionReader: Fetch chunk 0 failed.
org.apache.celeborn.common.exception.CelebornIOException: FETCH_DATA_TIMEOUT
at org.apache.celeborn.common.network.client.TransportResponseHandler.failExpiredFetchRequest(TransportResponseHandler.java:147)
at org.apache.celeborn.common.network.client.TransportResponseHandler.lambda$new$1(TransportResponseHandler.java:103)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
23/06/14 11:14:54 WARN RssInputStream: Fetch chunk failed 1/6 times for location PartitionLocation[
id-epoch:35-0
host-rpcPort-pushPort-fetchPort-replicatePort:10.169.48.203-9092-9094-9093-9095
mode:MASTER
peer:(host-rpcPort-pushPort-fetchPort-replicatePort:10.169.48.202-9092-9094-9093-9095)
storage hint:StorageInfo{type=HDD, mountPoint='/mnt/ssd/0', finalResult=true, filePath=}
mapIdBitMap:null], change to peer
org.apache.celeborn.common.exception.CelebornIOException: Fetch chunk 0 failed.
at org.apache.celeborn.client.read.WorkerPartitionReader$1.onFailure(WorkerPartitionReader.java:98)
at org.apache.celeborn.common.network.client.TransportResponseHandler.failExpiredFetchRequest(TransportResponseHandler.java:146)
at org.apache.celeborn.common.network.client.TransportResponseHandler.lambda$new$1(TransportResponseHandler.java:103)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: org.apache.celeborn.common.exception.CelebornIOException: FETCH_DATA_TIMEOUT
at org.apache.celeborn.common.network.client.TransportResponseHandler.failExpiredFetchRequest(TransportResponseHandler.java:147)
... 8 more
23/06/14 11:14:54 INFO SortBasedShuffleWriter: Memory used 72.0 MB
```
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Closes#1587 from AngersZhuuuu/CELEBORN-676.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
### What changes were proposed in this pull request?
CommitHandler will check whether the target worker is in WorkerStatusTracker's excluded list. If so, skip calling commit files on it.
### Why are the changes needed?
Avoid unnecessary commit files to excluded worker.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manual test.
Closes#1581 from waitinfuture/669.
Lead-authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Co-authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Co-authored-by: Keyong Zhou <zhouky@apache.org>
Signed-off-by: Shuang <lvshuang.tb@gmail.com>
### What changes were proposed in this pull request?
In our prod meet many times of push queue stuck caused by PushState's status was not being removed.
Caused DataPushQueue to keep waiting for taking task.
Although have resolved some bugs, here we'd better add a max wait time for taking tasks since we already have the `PUSH_DATA_TIMEOUT` check method. If the target worker is really stuck, we can retry our task finally.
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Closes#1552 from AngersZhuuuu/CELEBORN-640.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
### What changes were proposed in this pull request?
Pluginconf might be hard to understand why Celeborn needs to config class.
### Why are the changes needed?
Ditto.
### Does this PR introduce _any_ user-facing change?
NO.
### How was this patch tested?
UT.
Closes#1524 from FMX/CELEBORN-610.
Authored-by: Ethan Feng <ethanfeng@apache.org>
Signed-off-by: Ethan Feng <ethanfeng@apache.org>
