From f31bf268a236f4ec16ea6af75a2200e65d3e68fb Mon Sep 17 00:00:00 2001
From: groot <yihua.mo@zilliz.com>
Date: Fri, 9 Apr 2021 15:28:51 +0800
Subject: [PATCH] fix failure to open vector index in mishards (#4910)

* fix #4778

Signed-off-by: yhmo <yihua.mo@zilliz.com>
---
 CHANGELOG.md                           |  1 +
 core/build-support/lint_exclusions.txt |  1 +
 core/src/db/DBImpl.cpp                 |  7 +++++++
 core/src/scheduler/task/SearchTask.cpp | 12 ++++++------
 4 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 21a8720adc..610bb3e8b4 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,7 @@ Please mark all change in change log and use the issue from GitHub
 
 # Milvus 1.1.0 (TBD)
 ## Bug
+-   \#4778 Fail to open vector index
 -   \#4858 Fix the crash when query for a large topk on GPU Flat
 -   \#4862 The inserted data number increase automatically without any operations
 -   \#4894 The capacity of bloom filter should be determined by the row count of its segment
diff --git a/core/build-support/lint_exclusions.txt b/core/build-support/lint_exclusions.txt
index 36a872f912..cbba7b3100 100644
--- a/core/build-support/lint_exclusions.txt
+++ b/core/build-support/lint_exclusions.txt
@@ -1,6 +1,7 @@
 *cmake-build-debug*
 *cmake-build-release*
 *cmake_build*
+*build*
 *src/index/thirdparty*
 *thirdparty*
 *easylogging++*
diff --git a/core/src/db/DBImpl.cpp b/core/src/db/DBImpl.cpp
index 5693a98bde..ce6858d1f1 100644
--- a/core/src/db/DBImpl.cpp
+++ b/core/src/db/DBImpl.cpp
@@ -1927,6 +1927,13 @@ DBImpl::QueryAsync(const std::shared_ptr<server::Context>& context, meta::FilesH
     LOG_ENGINE_DEBUG_ << LogOut("Engine query begin, index file count: %ld", files.size());
     scheduler::SearchJobPtr job = std::make_shared<scheduler::SearchJob>(tracer.Context(), k, extra_params, vectors);
     for (auto& file : files) {
+        // no need to process shadow files
+        if (file.file_type_ == milvus::engine::meta::SegmentSchema::FILE_TYPE::NEW ||
+            file.file_type_ == milvus::engine::meta::SegmentSchema::FILE_TYPE::NEW_MERGE ||
+            file.file_type_ == milvus::engine::meta::SegmentSchema::FILE_TYPE::NEW_INDEX) {
+            continue;
+        }
+
         scheduler::SegmentSchemaPtr file_ptr = std::make_shared<meta::SegmentSchema>(file);
         job->AddIndexFile(file_ptr);
     }
diff --git a/core/src/scheduler/task/SearchTask.cpp b/core/src/scheduler/task/SearchTask.cpp
index f04c1aed61..b74ef08c8c 100644
--- a/core/src/scheduler/task/SearchTask.cpp
+++ b/core/src/scheduler/task/SearchTask.cpp
@@ -111,14 +111,14 @@ XSearchTask::XSearchTask(const std::shared_ptr<server::Context>& context, Segmen
             ascending_reduce = false;
         }
 
-        EngineType engine_type;
-        if (file->file_type_ == SegmentSchema::FILE_TYPE::RAW ||
-            file->file_type_ == SegmentSchema::FILE_TYPE::TO_INDEX ||
-            file->file_type_ == SegmentSchema::FILE_TYPE::BACKUP) {
+        // The file may be a raw file, or an index file.
+        // Deduce the file's engine_type by this rule:
+        //     If the file is a raw file, its file_id is equal to segment id, engine_type is IDMAP
+        //     else if the file is an index file, its file_id is different to segment id, engine_type is index_type
+        EngineType engine_type = (EngineType)file->engine_type_;
+        if (file->segment_id_ == file->file_id_) {
             engine_type = engine::utils::IsBinaryMetricType(file->metric_type_) ? EngineType::FAISS_BIN_IDMAP
                                                                                 : EngineType::FAISS_IDMAP;
-        } else {
-            engine_type = (EngineType)file->engine_type_;
         }
 
         milvus::json json_params;