Resume CodeQL database creation after Java OOM at finalizing stage

[hx1997]

Hi there,

I'm trying to create a database for the Android AOSP. The building completed successfully but at the last moment it told me there was not enough Java heap space (see log below). Is there a way I can resume from the finalization stage (as opposed to starting all over) after I increase my memory? As you see, the building takes a lot of time and it'd be a pain in the neck to have to do it again from the beginning.

I tried rerunning the database create command, but it complained:

A fatal error occurred: Refusing to create database: '/mnt/g/aosp/out/codeql-bt' exists and is not an empty directory.
Please specify a different location, or use the --overwrite flag to overwrite existing CodeQL databases.

When I added the --overwrite flag, the message seemed to suggest me to delete it all (!) and start over:

A fatal error occurred: Ran with database overwriting enabled, but the directory does not appear to be a CodeQL database or database cluster. Please check you do indeed wish to delete it, and do so manually.

Thanks.

CodeQL Log:

[2023-01-29 16:21:57] [build-stdout] #### build completed successfully (52:10:24 (hh:mm:ss)) ####
Finalizing database at /mnt/g/aosp/out/codeql-bt.
Oops! A fatal internal error occurred.
java.lang.RuntimeException: java.lang.OutOfMemoryError: Java heap space
        at com.semmle.util.exception.Exceptions.asUnchecked(Exceptions.java:103)
        at com.semmle.inmemory.trap.TrapImporter.run(TrapImporter.java:111)
        at com.semmle.cli2.ql.dataset.ImportCommand.executeSubcommand(ImportCommand.java:92)
        at com.semmle.cli2.picocli.SubcommandCommon.lambda$executeSubcommandWithMessages$2(SubcommandCommon.java:625)
        at com.semmle.cli2.picocli.SubcommandCommon.withCompilationMessages(SubcommandCommon.java:405)
        at com.semmle.cli2.picocli.SubcommandCommon.executeSubcommandWithMessages(SubcommandCommon.java:624)
        at com.semmle.cli2.picocli.PlumbingRunner.run(PlumbingRunner.java:109)
        at com.semmle.cli2.picocli.SubcommandCommon.runPlumbingInProcess(SubcommandCommon.java:188)
        at com.semmle.cli2.database.FinalizeCommand.doTrapImport(FinalizeCommand.java:182)
        at com.semmle.cli2.database.FinalizeCommand.finalizeOne(FinalizeCommand.java:147)
        at com.semmle.cli2.database.FinalizeCommand.executeSubcommand(FinalizeCommand.java:90)
        at com.semmle.cli2.picocli.SubcommandCommon.lambda$executeSubcommandWithMessages$2(SubcommandCommon.java:625)
        at com.semmle.cli2.picocli.SubcommandCommon.withCompilationMessages(SubcommandCommon.java:405)
        at com.semmle.cli2.picocli.SubcommandCommon.executeSubcommandWithMessages(SubcommandCommon.java:624)
        at com.semmle.cli2.picocli.PlumbingRunner.run(PlumbingRunner.java:109)
        at com.semmle.cli2.picocli.SubcommandCommon.runPlumbingInProcess(SubcommandCommon.java:188)
        at com.semmle.cli2.database.CreateCommand.executeSubcommand(CreateCommand.java:169)
        at com.semmle.cli2.picocli.SubcommandCommon.lambda$executeSubcommandWithMessages$2(SubcommandCommon.java:625)
        at com.semmle.cli2.picocli.SubcommandCommon.withCompilationMessages(SubcommandCommon.java:405)
        at com.semmle.cli2.picocli.SubcommandCommon.executeSubcommandWithMessages(SubcommandCommon.java:624)
        at com.semmle.cli2.picocli.SubcommandCommon.call(SubcommandCommon.java:561)
        at com.semmle.cli2.picocli.SubcommandMaker.runMain(SubcommandMaker.java:225)
        at com.semmle.cli2.picocli.SubcommandMaker.runMain(SubcommandMaker.java:234)
        at com.semmle.cli2.CodeQL.main(CodeQL.java:100)
Caused by: java.lang.OutOfMemoryError: Java heap space
        at java.base/java.lang.StringUTF16.compress(Unknown Source)
        at java.base/java.lang.String.<init>(Unknown Source)
        at java.base/java.lang.String.<init>(Unknown Source)
        at java.base/java.io.BufferedReader.readLine(Unknown Source)
        at java.base/java.io.BufferedReader.readLine(Unknown Source)
        at com.semmle.util.trap.dependencies.TextFile.load(TextFile.java:59)
        at com.semmle.util.trap.dependencies.TrapSet.<init>(TrapSet.java:59)
        at com.semmle.inmemory.trap.TRAPLinker.addTrapAndDependenciesToSet(TRAPLinker.java:439)
        at com.semmle.inmemory.trap.TRAPLinker.computeObjectToTrap(TRAPLinker.java:423)
        at com.semmle.inmemory.trap.TRAPLinker.getTasks(TRAPLinker.java:179)
        at com.semmle.inmemory.trap.TrapImporter.lambda$scanAndLink$3(TrapImporter.java:185)
        at com.semmle.inmemory.trap.TrapImporter$$Lambda$356/0x0000000800312ca8.apply(Unknown Source)
        at com.semmle.util.concurrent.FutureUtils.lambda$mapAsync$4(FutureUtils.java:120)
        at com.semmle.util.concurrent.FutureUtils$$Lambda$358/0x0000000800313340.get(Unknown Source)
        at java.base/java.util.concurrent.CompletableFuture$AsyncSupply.run(Unknown Source)
        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)

[aibaars]

The codeql database create command is actually composed of 3 subcommands: database init, database trace-command and database finalize . You should be able to run codeql database finalize to finish the database creation.

[hx1997]

Thanks! I tried to run codeql database finalize -v -- out/codeql-bt/ following your suggestion, but it told me the directory is not a CodeQL database: A fatal error occurred: /mnt/g/aosp/out/codeql-bt is not a recognized CodeQL database.

[aibaars]

Oh, that is strange. Is there a codeql-database.yml file in /mnt/g/aosp/out/codeql-b ?

[hx1997]

No, there are only three subdirectories: log, src, and trap. Could it be my attempted rerunning of the database create command deleted it somehow?

[aibaars]

Yes, I think so.

[hx1997]

How unfortunate. Maybe a warning should be added to database create against that. I remember now: I reran the create command with --overwrite flag twice. The first time it said I had to specify --overwrite. I was baffled and ran it a second time, which displayed the "not a CodeQL database" error. The .yml file must have been deleted in the first run. No extra warning is needed because it's me who explicitly instructed CodeQL to overwrite the database. :(