Bridge server is unresponsive

SonarQube Cloud
1
  • ALM used: Azure DevOps
  • CI :Azure DevOps
  • Scanner command used when applicable: SonarCloudAnalyze@2 & SonarCloudPublish@2
  • Languages of the repository: C#
  • Only if the SonarCloud project is public, the URL: SonarQube Cloud
  • Error observed:

We’re getting an error stating that the “bridge server is unresponsive”. The issue has persisted for the better part of the last 2 days.

  • Steps to reproduce: Run any pipeline containing the Sonar steps.
  • Potential workaround: Disabling sonar, but that’s not the point
2

Hi,

Welcome to the community!

Can you try increasing sonar.javascript.node.maxspace?

 
Ann

3

Hi, how do I do that on Azure DevOps? I don’t think the tasks have this documented? What value should I increase it to? The VM running the build is a ~14GB VM

4

Hi,

Did you check the docs link in my previous post?

 
Ann

5

Ah, there it is.

Some followup questions:

  • What is the default setting?
  • Would you recommend using different settings for .NET vs NodeJS projects?
6

Hi,

Well for a pure .NET project you wouldn’t need this at all :joy:

More seriously, this is not something I would bump proactively on most projects unless you know going in that you’ll have a lot going on with JS/TS analysis.

 
HTH,
Ann

7

Yeah it’s related to a JS/TS project. What is the default value being used btw?

8

Hi,

I’m not sure off-hand. The docs recommend

Set this property to 4096 or 8192 for big projects.

Given that doubling, I would guess the default is 2048, but that’s only a guess.

 
HTH,
Ann

9

We’ve tried setting the values you’ve prescribed, and changed the build agent vm to a 3x core / 14GB RAM image. It didn’t help. The issue persists.

I’m going through the docs again. Is there anything else you could recommend?

10

Hi,

Can you share the full analysis log?

The analysis / scanner log is what’s output from the analysis command. Hopefully, the log you provide - redacted as necessary - will include that command as well.

This guide will help you find them.

 
Thx,
Ann

11

I’ll get to that today. I’ll update when I have the log

1 Like
12

Hello, I’m attaching the log from a recent execution, as well as the SonarCloudPrepare step which is running prior to the analysis.

  - task: SonarCloudPrepare@3
        inputs:
          SonarCloud: <redacted>
          organization: <redacted>
          scannerMode: cli
          configMode: manual
          cliProjectKey: $(Build.Repository.Name)
          cliSources: src
          extraProperties: |
            sonar.verbose=true
            sonar.exclusions=**/node_modules/**
            sonar.test.inclusions=**/*.(spec|test).*
            sonar.coverage.exclusions=**/*.(spec|test).*
            sonar.javascript.lcov.reportPaths=$(lcovReportPath)
            sonar.javascript.node.maxspace=8192
        displayName: Sonar config

sonar.log (1.3 MB)

13

Hi,

Thanks for the log.

1525 files indexed

Your project is really big. You may need to just keep doubling the memory allocation until it works.

 
Ann

14

I’ve noticed that the scanner and JVM use 3GB RAM each. Is this a potential factor?

Also, the scanner suddenly stopped working after a developer added ~10 LoC in a PullRequest. Up until then it was working fine under the default settings. I’m not entirely sure if the doubling strategy is the correct way to go here

15

Hi,

What were those 10 LoC? Were they particularly complex?

 
Ann

16

No, not really.

One class was added to an import statement in TS, and some simple business logic was modified.

17

Hi,

It may be that you hit a tipping point. Have you tried the doubling strategy?

 
Ann

18

Our agent is a 14GB VM from the Azure Pipelines macos-latest pool.
If we double 8192 then we will exceed it’s memory limit.

At the moment, it appears that running the analysis utilizes 10GB out of the 14GB RAM, irregardless of the sonar.javascript.node.maxspace setting.

Our current configuration is:

      - task: SonarCloudPrepare@3
        inputs:
          SonarCloud: <redacted>
          organization: <redacted>
          scannerMode: cli
          configMode: manual
          cliProjectKey: $(Build.Repository.Name)
          cliSources: src
          extraProperties: |
            sonar.sources=src
            sonar.verbose=true
            sonar.sourceEncoding=UTF-8
            sonar.javascript.lcov.reportPaths=$(lcovReportPath)
            sonar.javascript.node.maxspace=4096
            sonar.inclusions=**/*.ts,**/*.tsx
sonar.exclusions=**/node_modules/**,**/dist/**,**/*.test.ts,**/*.test.tsx,**/__tests__/**

We’ve managed to get the indexed files down to 1069

19

screenshot
screenshot1256×1053 255 KB

This is running under sonar.javascript.node.maxspace=4096

20

This is during a PR analysis. Interestingly it gets stuck at the 1st file - which was the file modified during the PR