[BUG]throw an err :Navigating frame was detached in aws lambda

[Forktechai]

chromium Version:116
puppeteer / puppeteer-core Version:21.1.0
Node.js Version:16
Lambda / GCF Runtime: public.ecr.aws/lambda/nodejs:16

page.goto(url, { waitUntil: 'networkidle0' }) threw the following exception:

2023-12-20T01:39:55.514Z 4e30b7ef-9859-4639-a518-b05b3e9cf52f ERROR Error: Navigating frame was detached
at LifecycleWatcher.#onFrameDetached (/var/task/node_modules/.pnpm/[email protected]/node_modules/puppeteer-core/lib/cjs/puppeteer/common/LifecycleWatcher.js:106:47)
at /var/task/node_modules/.pnpm/[email protected]/node_modules/puppeteer-core/lib/cjs/third_party/mitt/index.js:1:269
at Array.map ()
at Object.emit (/var/task/node_modules/.pnpm/[email protected]/node_modules/puppeteer-core/lib/cjs/third_party/mitt/index.js:1:252)
at Frame.emit (/var/task/node_modules/.pnpm/[email protected]/node_modules/puppeteer-core/lib/cjs/puppeteer/common/EventEmitter.js:88:22)
at FrameManager.#removeFramesRecursively (/var/task/node_modules/.pnpm/[email protected]/node_modules/puppeteer-core/lib/cjs/puppeteer/common/FrameManager.js:394:15)
at /var/task/node_modules/.pnpm/[email protected]/node_modules/puppeteer-core/lib/cjs/puppeteer/common/FrameManager.js:90:46
at onceHandler (/var/task/node_modules/.pnpm/[email protected]/node_modules/puppeteer-core/lib/cjs/puppeteer/common/EventEmitter.js:99:13)
at /var/task/node_modules/.pnpm/[email protected]/node_modules/puppeteer-core/lib/cjs/third_party/mitt/index.js:1:269
at Array.map ()

but when i use the base image public.ecr.aws/lambda/nodejs:16.2023.11.10.17 , it works well.

[Sparticuz]

You'll need to update to 119.0.2 to get the new libs package.

[Prinzhorn]

I'm just in the process of migrating from an ancient chrome-aws-lambda and seeing the same error with page.goto. I tried @sparticuz/[email protected] with both [email protected] and [email protected]. According to https://pptr.dev/chromium-support one of them should definitely work or no?

[steadweb]

I've also seen this issue, we're currently seeing issues with libnss3 which may have been updated in the latest node:16 AWS lambda runtime image.

If you're bundling your own code in a docker image and running that on lambda, can you try the following in your Dockerfile:

FROM public.ecr.aws/lambda/nodejs:16.2023.11.18.13

Reverting back to an older node:16 version that hasn't updated libnss3 for us, works.

[rdsedmundo]

We're also seeing the same with latest @sparticuz/chromium and puppeteer-core.

[yoeven]

I had the same issue and have been on it for the entire day. it seems to be something related with the browser data being created in the /tmp directory being shared between lambda function. await browser.close() would technically clean up any data without affecting the next run but calling await browser.close() stalls sometimes. Here are something I did to fix it:

• Upgrade puppeteer-core to "puppeteer-core": "^21.10.0"
• Upgraded @sparticuz/chromium to "@sparticuz/chromium": "^121.0.0"
• call await browser.close() at the end of the function like this await Promise.race([browser.close(), browser.close(), browser.close()]); so that if one call hangs most of the time the others would work to continue the process
• remember to call await browser.close() even when your script throws errors for another reason, always close the browser before responding with a status

[alexander-dermicus]

Potentially related (with fix): #229 (comment)