Error initializing IMB disk cache results in no results

首先在 CodeQL extension 中重现重启序列：运行一个查询，更改一个会重启查询服务器的设置，然后再次运行该查询。跟踪堆栈跟踪中列出的 QueryServerBackendManager 和 EvaluationServer 入口点，并验证已停止的查询会释放 IMB cache lock。后续查询能够成功执行而不是返回无结果，即表示完成。

Describe the bug
Changing the settings of the CodeQL extension while running a query results in the restart of the query server.
This restart successfully stops the running query, but is unable to properly release the lock on the IMB cache directory of the selected database.

Subsequent query executions return with a no result response and the query log contains the stack trace (with *** if information is redacted)

Jun 29, 2021 1:25:48 PM org.eclipse.lsp4j.jsonrpc.RemoteEndpoint fallbackResponseError
SEVERE: Internal error: Error initializing the IMB disk cache: the cache directory is already locked by another running process. Only one instance of the IMB can access a cache directory at a time. The lock file is located at /Users/***/Library/Application Support/Code/User/workspaceStorage/***/GitHub.vscode-codeql/***/***/db-cpp/default/cache/.lock
(eventual cause: OverlappingFileLockException)
java.util.concurrent.CompletionException: Error initializing the IMB disk cache: the cache directory is already locked by another running process. Only one instance of the IMB can access a cache directory at a time. The lock file is located at /Users/***/Library/Application Support/Code/User/workspaceStorage/***/GitHub.vscode-codeql/***/***/db-cpp/default/cache/.lock
(eventual cause: OverlappingFileLockException)
	at java.base/java.util.concurrent.CompletableFuture.encodeThrowable(Unknown Source)
	at java.base/java.util.concurrent.CompletableFuture.uniComposeStage(Unknown Source)
	at java.base/java.util.concurrent.CompletableFuture.thenCompose(Unknown Source)
	at com.semmle.api.server.QueryServerBackendManager.lambda$withDB$0(QueryServerBackendManager.java:70)
	at com.semmle.util.concurrent.FutureUtils.supplyCompose(FutureUtils.java:217)
	at com.semmle.util.concurrent.FutureUtils.futureFinally(FutureUtils.java:159)
	at com.semmle.api.server.QueryServerBackendManager.withDB(QueryServerBackendManager.java:70)
	at com.semmle.api.server.QueryServerBackendManager.withDBExclusive(QueryServerBackendManager.java:57)
	at com.semmle.api.server.EvaluationServer.lambda$clearCache$11(EvaluationServer.java:253)
	at com.semmle.api.server.QueryServerProgressHelper.lambda$null$0(QueryServerProgressHelper.java:47)
	at com.semmle.util.concurrent.FutureUtils.supplyCompose(FutureUtils.java:217)
	at com.semmle.api.server.QueryServerProgressHelper.lambda$withProgressAndCancellation$1(QueryServerProgressHelper.java:47)
	at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source)
	at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
	at java.base/java.lang.Thread.run(Unknown Source)
Caused by: Error initializing the IMB disk cache: the cache directory is already locked by another running process. Only one instance of the IMB can access a cache directory at a time. The lock file is located at /Users/***/Library/Application Support/Code/User/workspaceStorage/***/GitHub.vscode-codeql/***/***/db-cpp/default/cache/.lock
(eventual cause: OverlappingFileLockException)
	at com.semmle.inmemory.caching.RelationCacheImpl.takeFilesystemLock(RelationCacheImpl.java:351)
	at com.semmle.inmemory.caching.RelationCacheImpl.create(RelationCacheImpl.java:110)
	at com.semmle.inmemory.caching.RelationManager.<init>(RelationManager.java:93)
	at com.semmle.inmemory.MemoryBackend.<init>(MemoryBackend.java:142)
	at com.semmle.api.server.QueryServerBackendManager.createBackend(QueryServerBackendManager.java:78)
	at com.semmle.api.server.QueryServerBackendManager$WaitingEntry.create(QueryServerBackendManager.java:99)
	at com.semmle.api.server.QueryServerBackendManager.scheduleUnblockedTasks(QueryServerBackendManager.java:139)
	at com.semmle.api.server.QueryServerBackendManager.withDB(QueryServerBackendManager.java:69)
	... 8 more
Caused by: java.nio.channels.OverlappingFileLockException
	at com.semmle.util.files.FileSystemLock.<init>(FileSystemLock.java:75)
	at com.semmle.inmemory.caching.RelationCacheImpl.takeFilesystemLock(RelationCacheImpl.java:347)
	... 15 more

Version
CodeQL CLI: v2.3.4
CodeQL Extension: v1.5.1

Visual Studio Code:
Version: 1.57.1 (Universal)
Commit: 507ce72a4466fbb27b715c3722558bb15afa9f48
Date: 2021-06-17T13:28:32.912Z
Electron: 12.0.7
Chrome: 89.0.4389.128
Node.js: 14.16.0
V8: 8.9.255.25-electron.0
OS: Darwin x64 20.5.0

To reproduce
Steps to reproduce the behavior.

1. Execute a query
2. Change settings in the CodeQL extension that restart the query server (e.g., the amount ram or nr of threads to use)
3. Re-execute the stopped query

Expected behavior
Execution of the query

Additional context
Add any other context about the problem here.