[improve] Refactor block file cache and add more ut case

be/src/common/config.cpp

@@ -984,20 +984,14 @@ DEFINE_Bool(enable_file_cache, "false");
 // format: [{"path":"/path/to/file_cache","total_size":21474836480,"query_limit":10737418240}]
 // format: [{"path":"/path/to/file_cache","total_size":21474836480,"query_limit":10737418240},{"path":"/path/to/file_cache2","total_size":21474836480,"query_limit":10737418240}]
 DEFINE_String(file_cache_path, "");
-DEFINE_Int64(file_cache_max_file_segment_size, "4194304"); // 4MB
-// 4KB <= file_cache_max_file_segment_size <= 256MB
-DEFINE_Validator(file_cache_max_file_segment_size, [](const int64_t config) -> bool {
-    return config >= 4096 && config <= 268435456;
-});
-DEFINE_Int64(file_cache_min_file_segment_size, "1048576"); // 1MB
-// 4KB <= file_cache_min_file_segment_size <= 256MB
-DEFINE_Validator(file_cache_min_file_segment_size, [](const int64_t config) -> bool {
-    return config >= 4096 && config <= 268435456 &&
-           config <= config::file_cache_max_file_segment_size;
-});
+DEFINE_Int64(file_cache_each_block_size, "1048576"); // 1MB
+// DEFINE_Validator(file_cache_each_block_size, [](const int64_t config) -> bool {
+//     return config <= config::s3_write_buffer_size && config::s3_write_buffer_size % config == 0;
+// });
 DEFINE_Bool(clear_file_cache, "false");
 DEFINE_Bool(enable_file_cache_query_limit, "false");
-DEFINE_mInt32(file_cache_wait_sec_after_fail, "0"); // // zero for no waiting and retrying
+DEFINE_mInt32(file_cache_enter_disk_resource_limit_mode_percent, "90");
+DEFINE_mInt32(file_cache_exit_disk_resource_limit_mode_percent, "80");
 
 DEFINE_mInt32(index_cache_entry_stay_time_after_lookup_s, "1800");
 DEFINE_mInt32(inverted_index_cache_stale_sweep_time_sec, "600");

be/src/common/config.h

@@ -1028,12 +1028,13 @@ DECLARE_Bool(enable_file_cache);
 // format: [{"path":"/path/to/file_cache","total_size":21474836480,"query_limit":10737418240},{"path":"/path/to/file_cache2","total_size":21474836480,"query_limit":10737418240}]
 // format: [{"path":"/path/to/file_cache","total_size":21474836480,"query_limit":10737418240,"normal_percent":85, "disposable_percent":10, "index_percent":5}]
 DECLARE_String(file_cache_path);
-DECLARE_Int64(file_cache_min_file_segment_size);
-DECLARE_Int64(file_cache_max_file_segment_size);
+DECLARE_Int64(file_cache_each_block_size);
 DECLARE_Bool(clear_file_cache);
 DECLARE_Bool(enable_file_cache_query_limit);
-// only for debug, will be removed after finding out the root cause
-DECLARE_mInt32(file_cache_wait_sec_after_fail); // zero for no waiting and retrying
+// default enter disk resource limit mode 5%
+DECLARE_Int32(file_cache_enter_disk_resource_limit_mode_percent);
+// default exit disk resource limit mode 20%
+DECLARE_Int32(file_cache_exit_disk_resource_limit_mode_percent);
 
 // inverted index searcher cache
 // cache entry stay time after lookup

be/src/io/cache/block/block_file_cache_settings.h → be/src/http/action/clear_file_cache_action.cpp

@@ -14,30 +14,27 @@
 // KIND, either express or implied.  See the License for the
 // specific language governing permissions and limitations
 // under the License.
-// This file is copied from
-// https://github.com/ClickHouse/ClickHouse/blob/master/src/Interpreters/Cache/FileCacheSettings.h
-// and modified by Doris
 
-#pragma once
+#include "http/action/clear_file_cache_action.h"
 
-#include <array>
+#include <fmt/core.h>
 
-#include "io/cache/block/block_file_cache_fwd.h"
+#include "http/http_channel.h"
+#include "http/http_request.h"
+#include "io/cache/block_file_cache_factory.h"
 
 namespace doris {
-namespace io {
 
-struct FileCacheSettings {
-    size_t total_size {0};
-    size_t disposable_queue_size {0};
-    size_t disposable_queue_elements {0};
-    size_t index_queue_size {0};
-    size_t index_queue_elements {0};
-    size_t query_queue_size {0};
-    size_t query_queue_elements {0};
-    size_t max_file_segment_size {0};
-    size_t max_query_cache_size {0};
-};
+const std::string SYNC = "sync";
+
+void ClearFileCacheAction::handle(HttpRequest* req) {
+    std::string sync = req->param(SYNC);
+    if (sync == "TRUE") {
+        io::FileCacheFactory::instance()->clear_file_caches(true);
+    } else {
+        io::FileCacheFactory::instance()->clear_file_caches(false);
+    }
+    HttpChannel::send_reply(req, HttpStatus::OK, Status::OK().to_json());
+}
 
-} // namespace io
 } // namespace doris

be/src/io/cache/block/block_file_cache_fwd.h → be/src/http/action/clear_file_cache_action.h

@@ -14,23 +14,19 @@
 // KIND, either express or implied.  See the License for the
 // specific language governing permissions and limitations
 // under the License.
-// This file is copied from
-// https://github.com/ClickHouse/ClickHouse/blob/master/src/Interpreters/Cache/FileCache_fwd.h
-// and modified by Doris
 
 #pragma once
-#include <memory>
 
-#include "vec/common/uint128.h"
+#include "http/http_handler.h"
 
-static constexpr size_t GB = 1 * 1024 * 1024 * 1024;
-static constexpr size_t KB = 1024;
 namespace doris {
-namespace io {
+class ExecEnv;
+class ClearFileCacheAction : public HttpHandler {
+public:
+    ClearFileCacheAction() = default;
 
-using uint128_t = vectorized::UInt128;
-using UInt128Hash = vectorized::UInt128Hash;
-static constexpr size_t REMOTE_FS_OBJECTS_CACHE_DEFAULT_ELEMENTS = 100 * 1024;
+    ~ClearFileCacheAction() override = default;
 
-} // namespace io
+    void handle(HttpRequest* req) override;
+};
 } // namespace doris

be/src/http/action/file_cache_action.cpp

@@ -26,7 +26,7 @@
 #include "http/http_headers.h"
 #include "http/http_request.h"
 #include "http/http_status.h"
-#include "io/cache/block/block_file_cache_factory.h"
+#include "io/cache/block_file_cache_factory.h"
 #include "olap/olap_define.h"
 #include "olap/tablet_meta.h"
 #include "util/easy_json.h"