This is an automated email from the ASF dual-hosted git repository. reswqa pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/flink.git
commit d8660309854d02a3f55e50f8d80b64c23bb76f55 Author: Zihao Chen <[email protected]> AuthorDate: Tue Jul 7 17:44:23 2026 +0800 [FLINK-40097][docs] Document archive load modes of HistoryServer --- .../docs/deployment/advanced/historyserver.md | 18 ++++++++++++++++++ docs/content/docs/deployment/advanced/historyserver.md | 18 ++++++++++++++++++ 2 files changed, 36 insertions(+) diff --git a/docs/content.zh/docs/deployment/advanced/historyserver.md b/docs/content.zh/docs/deployment/advanced/historyserver.md index 748dbd867fc..21ef3728c23 100644 --- a/docs/content.zh/docs/deployment/advanced/historyserver.md +++ b/docs/content.zh/docs/deployment/advanced/historyserver.md @@ -98,6 +98,24 @@ HistoryServer 支持通过 `historyserver.archive.storage.type` 选择本地存 historyserver.archive.storage.type: ROCKSDB ``` +**加载模式** + +HistoryServer 支持两种加载模式，通过 `historyserver.archive.load.mode` 配置： + +* `EAGER`（默认）：通过周期性的后台刷新，自动、同步地将存档下载到本地存储。 +* `LAZY`：存档会立即展示在 Web 界面上，底层数据则在后台异步拉取。如果某个存档在其后台下载完成之前被访问，系统会优先按需拉取该存档。 + +启用懒加载模式示例： + +```yaml +historyserver.archive.load.mode: LAZY +``` + +在 `LAZY` 模式下，按需拉取使用两个线程池： + +* `historyserver.lazy.fetch.executor.common.pool-size` —— 用于常规后台存档拉取的共享线程池的大小。 +* `historyserver.lazy.fetch.executor.individual.pool-size` —— 专用于按需拉取单个存档（例如用户访问某个存档时）的高优先级线程池的大小。 + ## 日志集成 Flink 本身并不提供已完成作业的日志收集功能。 diff --git a/docs/content/docs/deployment/advanced/historyserver.md b/docs/content/docs/deployment/advanced/historyserver.md index 607f08b826b..842b73e3a0b 100644 --- a/docs/content/docs/deployment/advanced/historyserver.md +++ b/docs/content/docs/deployment/advanced/historyserver.md @@ -92,6 +92,24 @@ Example for enabling the RocksDB backend: historyserver.archive.storage.type: ROCKSDB ``` +**Archive Load Mode** + +The HistoryServer supports two modes for loading archives, selected via `historyserver.archive.load.mode`: + +* `EAGER` (default): Archives are automatically and synchronously downloaded to local storage via periodic background refreshes. +* `LAZY`: Archives are displayed on the web interface immediately, while the underlying data is fetched asynchronously in the background. If a specific archive is accessed before its background download completes, the system will prioritize and fetch it on demand. + +Example for enabling the lazy loading: + +```yaml +historyserver.archive.load.mode: LAZY +``` + +In `LAZY` mode, on-demand fetches use two thread pools: + +* `historyserver.lazy.fetch.executor.common.pool-size` — The size of the shared thread pool used for routine, background archive fetching. +* `historyserver.lazy.fetch.executor.individual.pool-size` — The size of the high-priority thread pool dedicated to fetching individual archives on demand (e.g., when a user accesses them). + ## Log Integration Flink does not provide built-in methods for archiving logs of completed jobs.
