I’m having trouble starting the local environment....
# help
ö
I’m having trouble starting the local environment. It was working all fine till today. I’m not really sure what the problem is. I removed all of the stacks through the CloudFormation Dashboard, and tried to deploy again, but no good. So I tried to create a brand new SST app, but the same problem persists. I’m not able to deploy the debug stack on a fresh serverless stack app.
Copy code
$ npx create-serverless-stack@latest demo --language typescript --use-yarn

$ cd demo

$ aws-vault exec sg-omer -- npx sst start --stage dev --verbose
I don’t see any CloudFormation Stack in progress:
I even deleted the
CDKToolKit
stack. It is interesting that it does progress creating the
CDKToolKit
stack. But after that, it seems to be stuck. I don’t know if it is relevant, but it also states that “the ongoing operation is not stable”
The CDKToolKit Stack is created, but I don’t see any progress on the terminal output
Still…
Sorry for pinging but.. @Frank @thdxr
f
Hey @Ömer Toraman, can you send me ur
.build/sst-debug.json
?
ö
Sure
f
can u check what state is the
development-softgiving-debug-stack
stack in?
ö
None… It does not exist at all
f
yeah, from the debug log, it seems it’s taking forever to upload the Lambda zip files to S3
ö
Checked all of the statuses from the dropdown:
The S3 bucket list seems to be buggy as well, it is showing me previously deleted buckets:
f
If u look at ur debug log:
Copy code
[2022-04-27T23:45:58.256] [TRACE] core - deploy stack: run cdk deploy: stderr: [0%] start: Publishing bdb08b9984816ecf43301a0c6d8913e270ee2448940b6b95e430dd222a4d90be:609221238233-us-east-1
to
Copy code
[2022-04-27T23:56:10.948] [TRACE] core - deploy stack: run cdk deploy: stderr: [25%] success: Published bdb08b9984816ecf43301a0c6d8913e270ee2448940b6b95e430dd222a4d90be:609221238233-us-east-1
It took 11 minutes to upload a file that’s likely a few KB in size
ö
So do you think it is about the internet speed?
f
can u check ur aws health page https://health.aws.amazon.com/health/status
Can u try manually creating a bucket on S3 and upload a 1 MB file in there and see if it’s quick?
ö
Can’t see any issue on aws health
It uploaded quickly:
it does create the CDKToolKit pretty fast
f
I see.. looking at the debug log again, yeah, it seems the uploading was actually fast. This was slow:
Copy code
[2022-04-27T23:45:58.872] [TRACE] core - deploy stack: run cdk deploy: stderr: [90m[0%] check: Check <s3://cdk-hnb659fds-assets-609221238233-us-east-1/bdb08b9984816ecf43301a0c6d8913e270ee2448940b6b95e430dd222a4d90be.zip>[39m
and then
Copy code
[2022-04-27T23:56:10.957] [TRACE] core - deploy stack: run cdk deploy: stderr: [90m[25%] check: Check <s3://cdk-hnb659fds-assets-609221238233-us-east-1/0d89df84ea307d9cb0d70fbd76481e4e86f449d075a2a082cc2d323ee69a98e4.zip>[39m
It seems CDK is make some call to S3 to check certain things, and that’s taking long.
ö
It did create some of the stacks
but it is taking really too long
the process is still going on
f
hmm.. can you try running
yarn cdk deploy
and see if deploying using CDK directly without going through SST is fast?
Can you also send me ur latest sst-debug.json?
ö
it finally deployed
I guess it took more than an hour lol
Sure, DM’ing you
It successfully deployed in just a couple of minutes, after my SSO login was expired and I relogged in
f
Oh do u think it’s related to SSO?
ö
I have no idea, the only explanation I have… Didn’t change anything at all
Idk how it could relate to aws-vault or sso
f
hmm one more thing we can check. If you go into AWS CloudTrail, you can see if there are failed requests during the time when sst deploy was slow.
If there were anything wrong with the previous SSO session, you might see a series of Rate Exceeded errors. That’d cause the AWS SDK calls to keep retrying, which could lead to the slowness.
Just a guess.