Command died with <Signals.SIGPIPE: 13> message

[vladaurosh]

I am using SSM plugin as proxy command for SSH to execute commands on EC2 instances. This is mostly used in jenkins pipelines.

Latest version of SSM plugin (session-manager-plugin-1.2.398.0-1.x86_64) is running on Centos7, installed with official rpm.

Occasionally (pretty rarely), I get this in output:
Command '['session-manager-plugin', '{"SessionId": "xxxxx", "TokenValue": "xxxxxxxxx", "StreamUrl": "wss://ssmmessages.eu-west-1.amazonaws.com/v1/data-channel/xxxx?role=publish_subscribe&cell-number=xxxxx", "ResponseMetadata": {"RequestId": "xxxx", "HTTPStatusCode": 200, "HTTPHeaders": {"server": "Server", "date": "Tue, 10 Jan 2023 12:10:51 GMT", "content-type": "application/x-amz-json-1.1", "content-length": "1059", "connection": "keep-alive", "x-amzn-requestid": "xxxx"}, "RetryAttempts": 0}}', 'eu-west-1', 'StartSession', 'xxxx', '{"Target": "i-xxxxxx", "DocumentName": "AWS-StartSSHSession", "Parameters": {"portNumber": ["22"]}}', 'https://ssm.eu-west-1.amazonaws.com']' died with <Signals.SIGPIPE: 13>.

Command doesn't fail, just getting this message where http status code is 200.

Any idea what's causing this? Or any way to hide the message (since it reveals some sensitive data) if it's not causing any issue?

[BastienBenech]

Hi, I had a similar problem using the SSM plugin (1.2.463.0-1) as ProxyCommand for SSH to connect on EC2 instances.

To use ProxyCommand, i had to execute my session manager command inside quotes, like that :
ProxyCommand sh -c "aws ssm..."

I sometimes had to use other quotes inside these commands, it is possible that these quotes create execution problems.

---

Edit (21/06/23) :
I have the same message whithout using quotes, so that's not the origin of the problem.