### What changes were proposed in this pull request?
Remove unnecessary configuration `celeborn.client.flink.inputGate.minMemory` and `celeborn.client.flink.resultPartition.minMemory`.
### Why are the changes needed?
`celeborn.client.flink.inputGate.minMemory` and `celeborn.client.flink.resultPartition.minMemory` are configured as min memory reserved at present. Meanwhile, `celeborn.client.flink.inputGate.memory` should be at least `networkBufferSize * MIN_BUFFERS_PER_GATE` bytes, and `celeborn.client.flink.resultPartition.memory` should be at least `networkBufferSize * MIN_BUFFERS_PER_PARTITION` bytes. Therefore, `celeborn.client.flink.inputGate.minMemory` and `celeborn.client.flink.resultPartition.minMemory` are unnecessary configuration for `celeborn.client.flink.inputGate.memory` and `celeborn.client.flink.resultPartition.memory`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
`PluginSideConfSuiteJ#testCoalesce`
Closes#2433 from SteNicholas/CELEBORN-1362.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
Before, there is no http request spec likes query param, http method and response mediaType.
And for each api, a HttpEndpoint class is needed.
In this PR, we refine the code for http service and provide swagger ui.
Note that: This pr does not change the orignal api request and response behavior, including metrics APIs.
TODO:
1. define DTO
2. http request authentication
<img width="1900" alt="image" src="https://github.com/apache/incubator-celeborn/assets/6757692/7f8c2363-170d-4bdf-b2c9-74260e31d3e5">
<img width="1138" alt="image" src="https://github.com/apache/incubator-celeborn/assets/6757692/3ae6ec8e-00a8-475b-bb37-0329536185f6">
### Why are the changes needed?
To close CELEBORN-1317
### Does this PR introduce _any_ user-facing change?
The api is align with before.
### How was this patch tested?
UT.
Closes#2371 from turboFei/jetty.
Authored-by: Fei Wang <fwang12@ebay.com>
Signed-off-by: Shuang <lvshuang.xjs@alibaba-inc.com>
Add SSL related configs and support for `ReloadingX509TrustManager`, required for enabling SSL support.
Please see #2416 for the consolidated PR with all the changes for reference.
Introduces SSL related configs for enabling and configuring use of TLS.
Yes, introduces configs to control behavior of SSL
The overall PR #2411 (and this PR as well) passes all tests, this is specifically pulling out the `ReloadingX509TrustManager` and config related changes
Closes#2419 from mridulm/config-for-ssl.
Authored-by: Mridul Muralidharan <mridulatgmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
`AbstractRemoteShuffleResultPartitionFactory` removes the check of shuffle compression codec.
### Why are the changes needed?
`AbstractRemoteShuffleResultPartitionFactory` checks whether shuffle compression codec is LZ4 for Flink 1.14 and 1.15 version at present. Meanwhile, since Flink 1.17 version, ZSTD has already been supported. Therefore `AbstractRemoteShuffleResultPartitionFactory` should remove the check of shuffle compression codec for Flink 1.17 version and above, which is checked via the constructor of `BufferCompressor`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
- `RemoteShuffleResultPartitionFactorySuiteJ`
Closes#2414 from SteNicholas/CELEBORN-1357.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: Shuang <lvshuang.xjs@alibaba-inc.com>
### What changes were proposed in this pull request?
Currently, the Celeborn master calculates the estimatedPartitionSize based on the fileInfo committed by the application. This estimate is then used to allocate slots across all workers. However, this partition size may be too large or too small for Celeborn. For example, if an application commits a single file of 1TB to only one worker, using that partition size could result in all other workers having no available slots or only very small slots. To improve this, it would be better to implement a cap on the master's estimated partition size to prevent such imbalances.
### Why are the changes needed?
As title
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
UT
Closes#2412 from RexXiong/CELEBORN-1345.
Lead-authored-by: lvshuang.xjs <lvshuang.xjs@taobao.com>
Co-authored-by: Shuang <lvshuang.xjs@alibaba-inc.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
Introduce JVM profiling `JVMProfier` in Celeborn Worker using async-profiler to capture CPU and memory profiles.
### Why are the changes needed?
[async-profiler](https://github.com/async-profiler) is a sampling profiler for any JDK based on the HotSpot JVM that does not suffer from Safepoint bias problem. It has low overhead and doesn’t rely on JVMTI. It avoids the safepoint bias problem by using the `AsyncGetCallTrace` API provided by HotSpot JVM to profile the Java code paths, and Linux’s perf_events to profile the native code paths. It features HotSpot-specific APIs to collect stack traces and to track memory allocations.
The feature introduces a profier plugin that does not add any overhead unless enabled and can be configured to accept profiler arguments as a configuration parameter. It should support to turn profiling on/off, includes the jar/binaries needed for profiling.
Backport [[SPARK-46094] Support Executor JVM Profiling](https://github.com/apache/spark/pull/44021).
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Worker cluster test.
Closes#2409 from SteNicholas/CELEBORN-1299.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: Shuang <lvshuang.xjs@alibaba-inc.com>
### What changes were proposed in this pull request?
add a new parameter to cap the max memory can be used for sort writer buffer
### Why are the changes needed?
with a huge number of partitions, the threshold based on buffer size * number of partitions without this cap can be too large, e.g. 64K * 100000 = 6G
### Does this PR introduce _any_ user-facing change?
a new parameter
### How was this patch tested?
ut
Closes#2388 from CodingCat/adaptive_followup.
Lead-authored-by: CodingCat <zhunansjtu@gmail.com>
Co-authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Co-authored-by: Keyong Zhou <zhouky@apache.org>
Co-authored-by: Keyong Zhou <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Improve Celeborn document to fix typos, formats, unvalid link and unsynced default value of document. Meanwhile, the public interfaces of `shuffleclient.md` keep the consistent with `ShuffleClient`.
### Why are the changes needed?
There are some typos, formats, unvalid link and unsynced default value fixes in Celeborn document at present. Meanwhile, the public interfaces of `shuffleclient.md` is inconsistent with `ShuffleClient`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
No.
Closes#2410 from SteNicholas/CELEBORN-1341.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
### Why are the changes needed?
`CELEBORN-1320` uses `ReviveManager` to batch processing SOFT_SPLIT event RPC, so `partitionSplitPool` is no longer used, and the configuration item `celeborn.client.push.splitPartition.threads` is meaningless.
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Closes#2396 from cxzl25/CELEBORN-1336.
Authored-by: sychen <sychen@ctrip.com>
Signed-off-by: SteNicholas <programgeek@163.com>
### What changes were proposed in this pull request?
Fix typo of `celeborn.network.bind.preferIpAddress` doc from `ture` to `true`.
### Why are the changes needed?
`celeborn.network.bind.preferIpAddress` doc has typo for `ture`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
No.
Closes#2392 from SteNicholas/prefer-ip-address.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: SteNicholas <programgeek@163.com>
### What changes were proposed in this pull request?
while SortBasedWriter has less memory footprint than HashBasedWriter, it suffers from performance issue when we have many partitions and the write buffer is filled with small chunks of data quickly
for example, if sort buffer size is 32K, you have 4 partitions and 128K data in total, the data distribution is like partition A, B, C, D, each time it comes with 8K per partition.... in this case, you need to compress and send small 8K chunk 4 times per partition , the cost would become very high. If you use hashbasedwriter, it doesn't have this problem since the push only happens when the per-partition buffer is full. Of course , larger sort buffer size can mitigate the issue, but tuning sort buffer size per job is a tedious work
this PR introduces a new feature that we measure total size of pushed bytes and pushed count as well as the "should-pushed" bytes and counts (should-push means that , the data we pushed is larger than CLIENT_PUSH_BUFFER_MAX_SIZE (in another word, we will trigger a push even with hashbasedwriter in this case))
when actualPushedBytes/actualPushedCounts > (1 + Threshold) * (ShouldPushBytes/ShouldPushCounts), we will enlarge the sort buffer size by 1X to try to buffer more data before pushing (the max size of sortBuffer would be capped at # of partitions * CLIENT_PUSH_BUFFER_MAX_SIZE)
### Why are the changes needed?
to reduce perf cost in sortbased writer
### Does this PR introduce _any_ user-facing change?
no, but have 2 extra configurations
### How was this patch tested?
in prod of our company and also unit test
Closes#2358 from CodingCat/adaptive_memory_threshold.
Authored-by: CodingCat <zhunansjtu@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
This enables a Celeborn Worker to retrieve the application meta from the Master if it hasn't received the secret from the Master before the application attempts to connect to it. Additionally, the Celeborn Worker's SecretRegistry has been converted into an LRU cache to prevent unbounded growth of the registry.
### Why are the changes needed?
This is last change needed for Auth support in Celeborn (https://issues.apache.org/jira/browse/CELEBORN-1011)
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added UTs and part of a bigger change which will be tested end-to-end.
Closes#2363 from otterc/CELEBORN-1179.
Authored-by: Chandni Singh <singh.chandni@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
This enables Celeborn Master to persist application meta in Ratis and also push it to Celeborn Workers when it receives the requests for slots from the LifecycleManager.
### Why are the changes needed?
This change is required for adding authentication. ([CELEBORN-1011](https://issues.apache.org/jira/browse/CELEBORN-1011)).
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added some UTs.
Closes#2346 from otterc/CELEBORN-1234.
Authored-by: Chandni Singh <singh.chandni@gmail.com>
Signed-off-by: SteNicholas <programgeek@163.com>
### What changes were proposed in this pull request?
This adds an internal port and auth support to Celeborn Wokers.
1. Internal port is used by a worker to receive messages from Celeborn Master.
2. Authentication support for secure communication with clients. This change doesn't add the support in clients to communicate to the Workers securely. That will be in a future change.
This change targets just adding the port and auth support to Worker. The following items from the proposal are still pending:
- Persisting the app secrets in Ratis.
- Forwarding secrets to Workers and having ability for the workers to pull registration info from the Master.
- Secured communication between workers and clients.
### Why are the changes needed?
It is needed for adding authentication support to Celeborn ([CELEBORN-1011](https://issues.apache.org/jira/browse/CELEBORN-1011))
### Does this PR introduce _any_ user-facing change?
Yes
### How was this patch tested?
Part of a bigger change. For this change, only modified existing UTs.
Closes#2292 from otterc/CELEBORN-1256.
Authored-by: Chandni Singh <singh.chandni@gmail.com>
Signed-off-by: waitinfuture <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Deprecate `celeborn.quota.configuration.path` config. User `celeborn.dynamicConfig.store.fs.path` instead.
### Why are the changes needed?
`DefaultQuotaManager` is removed in #2298, which causes that `celeborn.quota.configuration.path` is useless. `celeborn.quota.configuration.path` could be deprecated that uses `celeborn.dynamicConfig.store.fs.path` to config quota.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
No.
Closes#2339 from SteNicholas/CELEBORN-1239.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: SteNicholas <programgeek@163.com>
### What changes were proposed in this pull request?
Introduce `celeborn.dynamicConfig.store.fs.path` config to configure the path of dynamic config file for fs store backend.
### Why are the changes needed?
`FsConfigServiceImpl` uses `celeborn.quota.configuration.path` to configure the path of dynamic config file for fs store backend at present. The path of dynamic config file should be introduced with `celeborn.dynamicConfig.store.fs.path` instead of quota configuration path.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
No.
Closes#2337 from SteNicholas/CELEBORN-1296.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: SteNicholas <programgeek@163.com>
### What changes were proposed in this pull request?
Add `celeborn.quota.enabled` at Master and Client side to enable checking quota
### Why are the changes needed?
`celeborn.quota.enabled` should be added in Master and Client side to enable quota check for Celeborn Master and Client.
### Does this PR introduce _any_ user-facing change?
Add categories of `celeborn.quota,enabled` with `master` and `client`.
### How was this patch tested?
No.
Closes#2318 from AngersZhuuuu/CELEBORN-1277.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
### What changes were proposed in this pull request?
https://github.com/apache/incubator-celeborn/pull/2292#discussion_r1497160753
Based on the above discussion, removing the additional secured port. The existing port will be used for secured communication when auth is enabled.
### Why are the changes needed?
These changes are for enabling authentication
### Does this PR introduce _any_ user-facing change?
Yes.
### How was this patch tested?
This removed additional secured port.
Closes#2327 from otterc/CELEBORN-1257.
Authored-by: Chandni Singh <singh.chandni@gmail.com>
Signed-off-by: waitinfuture <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Change the default value of `celeborn.worker.graceful.shutdown.recoverDbBackend` from `LEVELDB` to `ROCKSDB`.
### Why are the changes needed?
Because the LevelDB support will be removed, the default value of `celeborn.worker.graceful.shutdown.recoverDbBackend` could be changed to ROCKSDB instead of LEVELDB for preparation of LevelDB deprecation.
Backport:
[[SPARK-45351][CORE] Change spark.shuffle.service.db.backend default value to ROCKSDB](https://github.com/apache/spark/pull/43142)
[[SPARK-45413][CORE] Add warning for prepare drop LevelDB support](https://github.com/apache/spark/pull/43217)
### Does this PR introduce _any_ user-facing change?
The default value of `celeborn.worker.graceful.shutdown.recoverDbBackend` is changed from `LEVELDB` to `ROCKSDB`.
### How was this patch tested?
No.
Closes#2320 from SteNicholas/CELEBORN-1280.
Lead-authored-by: SteNicholas <programgeek@163.com>
Co-authored-by: Nicholas Jiang <programgeek@163.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
This pr does 2 things:
1. Remove unnecessary conf QUOTA_MANAGER since we implement it with ConfigService and ConfigService already have a conf to indicate the implement method.
2. Move the quota manager to Master side since only master use this
3. Support quota manager use FsConfigService and support default system level
### Why are the changes needed?
1. Many times, for users who do not have a quota configured, we hope to have a default quota that applies to them.
2. Quota manager should support refresh
3. QuotaManager should support integrate with ConfigService
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
added ut
Closes#2298 from AngersZhuuuu/CELEBORN-1239.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
### What changes were proposed in this pull request?
Rename `celeborn.worker.sortPartition.reservedMemory.enabled` to `celeborn.worker.sortPartition.prefetch.enabled`. Address [r1469066327](https://github.com/apache/incubator-celeborn/pull/2264/files#r1469066327) of pan3793.
### Why are the changes needed?
`celeborn.worker.sortPartition.reservedMemory.enabled` is misleading, which should represent that prefetch the original partition files during the first sequential reading path to leverage the Linux PageCache mechanism to speed up the subsequent random reading of them. The config name could use `celeborn.worker.sortPartition.prefetch.enabled` which is is more accurate.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
No.
Closes#2312 from SteNicholas/CELEBORN-1254.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: SteNicholas <programgeek@163.com>
### What changes were proposed in this pull request?
Since we support ConfigService, many configuration can be dynamic, add `isDynamic` property for CelebornConf in this pr.
### Why are the changes needed?
Make configuration doc more cleear
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existed UT
Closes#2308 from AngersZhuuuu/CELEBORN-1051.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
### What changes were proposed in this pull request?
Improve the implementation of `ConfigService` including:
- Removes `celeborn.dynamicConfig.enabled`.
- Changes `celeborn.dynamicConfig.store.backend` to optional.
- Renames `refreshAllCache` to `refreshCache` in `ConfigService`.
- Checks whether the dynamic config file exists and is file in `FsConfigServiceImpl`.
### Why are the changes needed?
Whether to enable dynamic config could check via whether `celeborn.dynamicConfig.store.backend` is provided, instead of `celeborn.dynamicConfig.enabled`. The `refreshAllCache` interface could rename to `refreshCache` and throw Exception simply. Meanwhile, `FsConfigServiceImpl` should check whether the dynamic config file exists and is file.
### Does this PR introduce _any_ user-facing change?
- Renames `refreshAllCache` to `refreshCache` in `ConfigService`.
### How was this patch tested?
- `ConfigServiceSuiteJ`
Closes#2304 from SteNicholas/CELEBORN-1052.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: SteNicholas <programgeek@163.com>
### What changes were proposed in this pull request?
As title.
### Why are the changes needed?
For some scenarios, if Celeborn cannot be used, users want to report an error directly instead of fallback.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
CI
Closes#2291 from kerwin-zk/add-config.
Authored-by: xiyu.zk <xiyu.zk@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
This adds a secured port to Celeborn Master which is used for secure communication with LifecycleManager.
This is part of adding authentication support in Celeborn (see CELEBORN-1011).
This change targets just adding the secured port to Master. The following items from the proposal are still pending:
1. Persisting the app secrets in Ratis.
2. Forwarding secrets to Workers and having ability for the workers to pull registration info from the Master.
3. Secured and internal port in Workers.
4. Secured communication between workers and clients.
In addition, since we are supporting both secured and unsecured communication for backward compatibility and seamless rolling upgrades, there is an additional change needed. An app which registers with the Master can try to talk to the workers on unsecured ports which is a security breach. So, the workers need to know whether an app registered with Master or not and for that Master has to propagate list of un-secured apps to Celeborn workers as well. We can discuss this more with https://issues.apache.org/jira/browse/CELEBORN-1261
### Why are the changes needed?
It is needed for adding authentication support to Celeborn (CELEBORN-1011)
### Does this PR introduce _any_ user-facing change?
Yes
### How was this patch tested?
Added a simple UT.
Closes#2281 from otterc/CELEBORN-1257.
Authored-by: Chandni Singh <singh.chandni@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Support database based store backend implementation for dynamic configuration management
### Why are the changes needed?
Currently celeborn provides `FsConfigServiceImpl` implementation for dynamic config service which is based on file system, We cloud Support database based store backend implementation.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
- `ConfigServiceSuiteJ#testDbConfig`
Closes#2273 from RexXiong/CELEBORN-1054.
Authored-by: Shuang <lvshuang.xjs@alibaba-inc.com>
Signed-off-by: Shuang <lvshuang.xjs@alibaba-inc.com>
### What changes were proposed in this pull request?
Introduce hot load for CelebornRackResolver.
### Why are the changes needed?
In production environment, we often expand the machine, so the rack configuration also needs to be updated in time.
### Does this PR introduce _any_ user-facing change?
master.md
### How was this patch tested?
UTs.
Closes#2246 from leixm/issue_1241.
Lead-authored-by: Xianming Lei <jerrylei@apache.org>
Co-authored-by: Xianming Lei <31424839+leixm@users.noreply.github.com>
Signed-off-by: Angerszhuuuu <angers.zhu@gmail.com>
### What changes were proposed in this pull request?
With authentication ([CELEBORN-1011](https://issues.apache.org/jira/browse/CELEBORN-1011)), the handling of messages from Client by Celeborn services (Master/workers) will go through a SASL handshake. However, messages exchanged between Masters and Workers will not.
With a single netty server on the Master/Workers handling both client and master/workers messages, differentiating between the two types of connection is a challenge. It will be better if Master/Workers have a separate designated port for Clients and a separate one for internal components (workers and other Masters).
In this change, we propose
- the config that enables creating dedicated internal ports on Masters/Workers.
- creation of the dedicated internal port in just the Master. A subsequent PR will add that creation of the dedicated internal port in Workers.
### Why are the changes needed?
This change is required for adding authentication. ([CELEBORN-1011](https://issues.apache.org/jira/browse/CELEBORN-1011)).
### Does this PR introduce _any_ user-facing change?
Yes, there are new configurations added.
### How was this patch tested?
Added a UT
Closes#2265 from otterc/CELEBORN-1012.
Authored-by: Chandni Singh <singh.chandni@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Introduce `celeborn.worker.sortPartition.reservedMemory.enabled` to support that `PartitionFilesSorter` seeks to position of each block and does not warmup for non-hdfs files.
### Why are the changes needed?
File sorting includes three steps: reading files, sorting MapIds, and writing files. The default block of Celeborn is 256k, and the number of blocks is about 1000, so the sorting process is very fast, and the main overhead is file reading and writing. There are roughly three options for the entire sorting process:
1. Memory of the file size is allocated in advance, the file is read in as a whole, MapId is parsed and sorted, and Blocks are written back to the disk in MapId order.
2. No memory is allocated, seek to the location of each block, parse and sort the MapId, and transfer the Blocks of the original file to the new file in the order of MapId.
3. Allocate a small block of memory (such as 256k), read the entire file sequentially, parse and sort the MapId, and transfer the block of the original file to the new file in the order of MapId.
From an IO perspective, at first glance, solution 1 uses sufficient memory and there is no sequential reading and writing; solution 2 has random reading and random writing; solution 3 has sequential writing. Intuitively solution 1 has better performance. Due to the existence of PageCache, when writing a file in solution 3, the original file is likely to be cached in PageCache. `PartitionFilesSorter` support solution3 with PageCache at present, which has better performance especially HDD disk. It's better to support solution2 with switch config that seeks to position of each block and does not warm up for non-hdfs files especially SDD disk.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
GA and cluster.
Closes#2264 from SteNicholas/CELEBORN-1254.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: SteNicholas <programgeek@163.com>
### What changes were proposed in this pull request?
Resource consumption of worker does not update when update interval of resource consumpution is greater than heartbeat interval.
<img width="1741" alt="截屏2024-01-24 14 49 50" src="https://github.com/apache/incubator-celeborn/assets/46485123/21cfd412-c69e-4955-8bc8-155ee470697d">
This pull request introduces below changes:
1. Avoid master repeat add gauge for same user
2. For worker, user resource consumption can directly get from worker's snapshot, didn't need update interval
### Why are the changes needed?
No.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
No.
Closes#2260 from AngersZhuuuu/CELEBORN-1252.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
Add configs' alternatives to doc.
### Why are the changes needed?
To help users use correct configs.
### Does this PR introduce _any_ user-facing change?
NO.
### How was this patch tested?
GA.
Closes#2253 from FMX/b1241.
Authored-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
`PushDataHandler` should build replicate factory to get client for sending replicate data instead of push client factory. Meanwhile, timeout checker of `TransportResponseHandler` should run with `replicate` module instead of `push`.
Follow up #2232.
### Why are the changes needed?
`PushDataHandler` uses push client factory to create client for replicating, which should use replicate factory, otherwise replicate module configuration does not take effect for replicating of worker server. Meanwhile, timeout checker of `TransportResponseHandler` runs with `push` module, which does not work well with replicate client for worker.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
GA and cluster.
Closes#2241 from SteNicholas/CELEBORN-1225.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
`PushDataHandler` should build replicate factory to get client for sending replicate data instead of push client factory.
### Why are the changes needed?
`PushDataHandler` uses push client factory to create client for replicating, which should use replicate factory, otherwise replicate module configuration does not take effect for replicating of worker server.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
GA and cluster.
Closes#2232 from SteNicholas/CELEBORN-1225.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
- Fix some typos.
### Why are the changes needed?
- Ditto.
### Does this PR introduce _any_ user-facing change?
- No.
### How was this patch tested?
- No need.
Closes#2214 from Radeity/fix-typo.
Authored-by: Aaron Wang <wangweirao16@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Revert "[CELEBORN-1150] support io encryption for spark".
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Closes#2208 from FMX/b1150-3.
Authored-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Add a cache in partition sorted and limit its max size.
### Why are the changes needed?
To reduce memory consumption in partition sort by tweak the index cache.
### Does this PR introduce _any_ user-facing change?
NO.
### How was this patch tested?
GA and cluster.
Closes#2194 from FMX/B1201.
Lead-authored-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
Co-authored-by: Keyong Zhou <zky.zhoukeyong@alibaba-inc.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Remove pipeline feature for sort based writer
### Why are the changes needed?
The pipeline feature is added as part of CELEBORN-295, for performance. Eventually, an unresolvable issue that would crash the JVM was identified in https://github.com/apache/incubator-celeborn/pull/1807, and after discussion, we decided to delete this feature.
### Does this PR introduce _any_ user-facing change?
No, the pipeline feature is disabled by default, there are no changes to users who use the default settings.
### How was this patch tested?
Pass GA.
Closes#2196 from pan3793/CELEBORN-891.
Authored-by: Cheng Pan <chengpan@apache.org>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
Dynamically determine the writing mode in Spark based on the number of partitions.
### Why are the changes needed?
Enhance the flexibility of shuffle writes to improve performance.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Add uts
Closes#2160 from lyy-pineapple/dynamic-write-mode.
Lead-authored-by: liangyongyuan <liangyongyuan@xiaomi.com>
Co-authored-by: cxzl25 <cxzl25@users.noreply.github.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
1. To support io encryption for spark.
### Why are the changes needed?
Ditto.
### Does this PR introduce _any_ user-facing change?
NO.
### How was this patch tested?
GA and manually test on a cluster.
Closes#2135 from FMX/B1150.
Authored-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
### What changes were proposed in this pull request?
Changes the version of the config to 0.5 given that 0.4 will be released soon.
### Why are the changes needed?
See above.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
NA
Closes#2165 from otterc/CELEBORN-1180.
Authored-by: Chandni Singh <singh.chandni@gmail.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?
One user reported that LifecycleManager's parmap can create huge number of threads and causes OOM.
There are four places where parmap is called:
1. When LifecycleManager commits files
2. When LifecycleManager reserves slots
3. When LifecycleManager setup connection to workers
4. When LifecycleManager call destroy slots
This PR fixes the fourth one. To be more detail, this PR eliminates `parmap` when destroying slots, and also replaces `askSync` with `ask`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manual test and GA.
Closes#2156 from waitinfuture/1167.
Lead-authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Co-authored-by: cxzl25 <cxzl25@users.noreply.github.com>
Co-authored-by: Keyong Zhou <waitinfuture@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
This adds the client side Sasl authentication support in the transport layer. Most of this code is taken from Apache Spark.
### Why are the changes needed?
The changes are needed for adding authentication to Celeborn. See [CELEBORN-1011](https://issues.apache.org/jira/browse/CELEBORN-1011).
### Does this PR introduce _any_ user-facing change?
Added a configuration for Sasl request timeout
### How was this patch tested?
Will be adding `CelebornSaslSuiteJ.java` (https://github.com/apache/incubator-celeborn/pull/2105) that tests the end-to-end Sasl flow.
Closes#2139 from otterc/CELEBORN-1157.
Authored-by: Chandni Singh <singh.chandni@gmail.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?
Since we are backporting #2145 to branch-0.3, and the configuration entry `celeborn.client.rpc.shared.threads` in #2145
has a start version of 0.4.0, this update aligns the version accordingly.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Pass GA
Closes#2153 from cfmcgrady/celeborn-1160-followup.
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?
As title
### Why are the changes needed?
One user reported that LifecycleManager's parmap can create huge number of threads and causes OOM.
There are four places where parmap is called:
1. When LifecycleManager commits files
2. When LifecycleManager reserves slots
3. When LifecycleManager setup connection to workers
4. When StorageManager calls close
This PR fixes the first one. To be more detail, this PR eliminates `parmap` when doing committing files, and also replaces `askSync` with `ask`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manual test and GA.
Closes#2145 from waitinfuture/1160.
Lead-authored-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
Co-authored-by: Keyong Zhou <waitinfuture@gmail.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
When request slots, filter workers excluded by application
### Why are the changes needed?
If worker alive but can not service, register shuffle will remove the worker from application client exclude list and next shuffle may reserve slots on this worker，this will cause application revive unexpectly
### Does this PR introduce _any_ user-facing change?
Yes, request slots will filter workers excluded by application
### How was this patch tested?
UT,
Closes#2131 from wangshengjie123/fix-request-slots-blacklist.
Authored-by: wangshengjie <wangshengjie3@xiaomi.com>
Signed-off-by: zky.zhoukeyong <zky.zhoukeyong@alibaba-inc.com>
### What changes were proposed in this pull request?
If the user does not use prometheus to collect monitoring metrics, but rather some other ones. Using metrics in JSON format would be more user-friendly.The PR supports JSON format for metrics.
### Why are the changes needed?
Ditto.
### Does this PR introduce _any_ user-facing change?
Metrics supports JSON format
### How was this patch tested?
Cluster test.
Closes#2089 from suizhe007/CELEBORN-1122.
Authored-by: qinrui <qr7972@gmail.com>
Signed-off-by: Shuang <lvshuang.tb@gmail.com>
### What changes were proposed in this pull request?
Follow up #2100. Mainly changes the package from scala to java of the codes in #2100. Meanwhile, `FsConfigServiceImpl#refresh` should directly return instead of refreshing configs.
### Why are the changes needed?
This PR follow up dynamic `ConfigService` at `SystemLevel` and `TenantLevel`, Dynamic configuration is a type of configuration that can be changed at runtime as needed in #2100. The implementation of `ConfigService` is based on Java codes, which are put into Scala package and cause that the spotless plugin does not format well. After the changes of the pull request, there are much code style changes generated from the package moving behavior.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
`ConfigServiceSuiteJ`.
Closes#2125 from SteNicholas/CELEBORN-1052.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: Shuang <lvshuang.tb@gmail.com>
### What changes were proposed in this pull request?
The `clientPushBufferMaxSize` config is also used by `CelebornInputStreamImpl`, it's a config about push side and should not be used by fetch side. This pr introduces a fetch config to replace it.
### Why are the changes needed?
As above
### Does this PR introduce _any_ user-facing change?
Yes, a new config `celeborn.client.fetch.buffer.size` is introduced.
### How was this patch tested?
Pass CI
Closes#2118 from exmy/celeborn-1145.
Authored-by: exmy <xumovens@gmail.com>
Signed-off-by: Cheng Pan <chengpan@apache.org>
### What changes were proposed in this pull request?
Introduce JVM monitoring in Celeborn Worker using JVMQuake to enable early detection of memory management issues and facilitate fast failure.
### Why are the changes needed?
When facing out-of-control memory management in Celeborn Worker we typically use JVMkill as a remedy by killing the process and generating a heap dump for post-analysis. However, even with jvmkill protection, we may still encounter issues caused by JVM running out of memory, such as repeated execution of Full GC without performing any useful work during the pause time. Since the JVM does not exhaust 100% of resources, JVMkill will not be triggered. Therefore JVMQuake is introduced to provide more granular monitoring of GC behavior, enabling early detection of memory management issues and facilitating fast failure. Refers to the principle of [jvmquake](https://github.com/Netflix-Skunkworks/jvmquake) which is a JVMTI agent that attaches to your JVM and automatically signals and kills it when the program has become unstable.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
`JVMQuakeSuite`
Closes#2061 from SteNicholas/CELEBORN-1092.
Authored-by: SteNicholas <programgeek@163.com>
Signed-off-by: Fu Chen <cfmcgrady@gmail.com>
