[BUG] Error: spawn ETXTBSY

[EugeneEl]

hi, I got very strange error, unfortunately I didn't find anything related to my issue.
If possible please have a look at this.

Environment

• chrome-aws-lambda Version: ^1.20.2,
• puppeteer / puppeteer-core Version: ^1.20.0
• OS:
• Node.js Version: 8 and 10

I tried to run it on different AWS node versions.

Logs:
2019-10-15T21:42:28.073Z 1930cd72-b389-41b5-b5bd-d4534dd0bc25 ERROR Error while generating pdf:Error: spawn ETXTBSY
2019-10-15T21:42:28.076Z 1930cd72-b389-41b5-b5bd-d4534dd0bc25 ERROR Error: params.Body is required
at ManagedUpload.validateBody (/var/task/node_modules/aws-sdk/lib/s3/managed_upload.js:271:13)
at ManagedUpload.configure (/var/task/node_modules/aws-sdk/lib/s3/managed_upload.js:123:10)
at new ManagedUpload (/var/task/node_modules/aws-sdk/lib/s3/managed_upload.js:95:10)
at features.constructor.upload (/var/task/node_modules/aws-sdk/lib/services/s3.js:1133:20)
at /var/task/services/awsUploader.js:18:8
at new Promise ()
at Object.uploadFile (/var/task/services/awsUploader.js:17:10)
at module.exports (/var/task/export.js:38:40)
2019-10-15T21:42:28.076Z 1930cd72-b389-41b5-b5bd-d4534dd0bc25 INFO finally:
2019-10-15T21:42:28.094Z 1930cd72-b389-41b5-b5bd-d4534dd0bc25 ERROR Invoke Error
{
"errorType": "Error",
"errorMessage": "params.Body is required",
"stack": [
"Error: params.Body is required",
" at ManagedUpload.validateBody (/var/task/node_modules/aws-sdk/lib/s3/managed_upload.js:271:13)",
" at ManagedUpload.configure (/var/task/node_modules/aws-sdk/lib/s3/managed_upload.js:123:10)",
" at new ManagedUpload (/var/task/node_modules/aws-sdk/lib/s3/managed_upload.js:95:10)",
" at features.constructor.upload (/var/task/node_modules/aws-sdk/lib/services/s3.js:1133:20)",
" at /var/task/services/awsUploader.js:18:8",
" at new Promise ()",
" at Object.uploadFile (/var/task/services/awsUploader.js:17:10)",
" at module.exports (/var/task/export.js:38:40)"
]
}

my code:

 const generatePDF = async data => {
   try {
     const browser = await chromium.puppeteer.launch({ 
      args: chromium.args,
      defaultViewport: chromium.defaultViewport,
       executablePath: await chromium.executablePath,
      headless: chromium.headless
     });
     const page = await browser.newPage();
 
     await page.goto(`file://${BASE_FILE}`, {
       waitUntil: 'networkidle0'
     });
 
     console.log(data);
 
     page.on('console', msg => console.log('PAGE LOG:', msg.text()));
 
     await page.evaluate(async params => await startVue(params), data);
 
     const pdf = await page.pdf({ format: 'A4' });
 
     browser.close();
     return pdf;
   } catch (error) {
     console.error(`Error while generating pdf:${error}`);
     throw error;
   }
 };

[alixaxel]

@EugeneEl It looks like the snippet you shared is incomplete, so hard to tell what's going on.

A couple of points:

• file:// the file protocol needs 3 forward slashes (like file:///...)
• what's BASE_FILE? does it exist on disk?
• what's startVue()? and why is it complaning that params.Body is required?

Please share the whole repro so that I can help you better.

[karimc1]

Hi Alixaxel,

I had the same error this is why I am leaving my comment on this closed issue:

When I deploy my lambda, I constantly get "spawn ETXTBSY" error.
When I delete the layer from AWS console and reattach it again, it works.
At the end of the lifecycle of the lambda (more than 10 minutes after), when calling the lambda again I fall back on the same error.
Do you have an idea?

Below the stack error as well as my generation function

{
	"errorType": "Error",
	"errorMessage": "spawn ETXTBSY",
	"trace": [
		"Error: spawn ETXTBSY",
		"    at ChildProcess.spawn (internal/child_process.js:366:11)",
		"    at Object.spawn (child_process.js:551:9)",
		"    at Launcher.launch (/var/task/node_modules/puppeteer-core/lib/Launcher.js:132:40)"
	]
}
​
const convertToPdf = async function (html) {
	let browser = null;
	let pdf = null;
	try {
		const executablePath = await chromium.executablePath;
		browser = await chromium.puppeteer.launch({
				args: chromium.args,
				defaultViewport: chromium.defaultViewport,
				executablePath: executablePath,
				headless: chromium.headless
			});
		const page = await browser.newPage();
		await page.setContent(html);
​
		pdf = await page.pdf({
				format: 'A4'
			});
​
	} catch (error) {
		throw error;
	}
	finally {
		if (browser !== null) {
			await browser.close();
		}
		if (pdf) {
			return pdf;
		}
	}
​
};

[alixaxel]

@karimc1 It's really hard to help without a complete repro. My guess is that you're calling convertToPdf multiple times from the same Lambda instance? Also, you only have a single chrome-aws-lambda layer attached right?

The ETXTBSY error code is thrown when puppeteer is trying to start the Chromium binary but the binary file is under a exclusive lock - that can happen if you don't wait for executablePath to resolve, or, if you call executablePath multiple times.

I'd be happy to help you further if you need, just open a new issue and share a full repro.

[karimc1]

Thanks for your quick answer I will open another ticket for that.

Bests,

[bwoodlt]

@karimc1 how did you fix it?

[krvajal]

I am having the same issue

[Eastkap]

having the issue too

[kryptonian41]

I am facing the same issue when launching chrome-aws-lambda browser inside WSL2 in headless mode. It used to work fine for quite a while now I don't whether I made a change in my environment which caused this.

[kryptonian41]

@alixaxel can you reopen the issue ?

[mikejackowski]

Sam issue happening to me, everything on the latest versions

[umairyounus]

Getting same issue

[SethCaryak]

I am having the same issue. First I get an error that it can't find chromium. {"errno":-2,"code":"ENOENT","syscall":"open","path":"/var/task/src/functions/bin/chromium.br"}"
subsequent calls result in "{"errno":-26,"code":"ETXTBSY","syscall":"spawn"}"
the chromium.executablepath is actually /tmp/chromium which I'm guessing is a symlink to /var/task/src/functions/bin/chromium.br which isn't actually there for some reason, the layer is setup correctly afaik. Been working on this for two days straight, need help, about to shoot myself

[jacksonneal]

I am running into this error when bundling with esbuild (only when in lambda, local invoke works fine). However, I can successfully execute when using webpack as in this repo https://github.com/metacollective9/playwright-amz-scrape

[gventuri]

Same happening for me, with node 16 and 12. Any idea?

[Carson12345]

@gventuri it could happen when you launch more than 1 puppeteer instance synchronously in your code.

is it your case?

[Moicky]

@gventuri it could happen when you launch more than 1 puppeteer instance synchronously in your code.

is it your case?

Well, this was the issue in my case 🤦

This comment made me read my code again and therefore I found the issue. Thanks a lot 😄

[AlteriusOmega]

Just in case anyone else made the same mistake as me, I forgot to comment out the running of the handler function at the end of my Lambda function that I use to run locally for testing during development. This caused two instances of the function to run causing this error.