parse AMI id from packer output

[Weng Wilson]

Hi,

I want to get AMI id (multiple AMIs from parallel building) from packer output and maybe record it to a json-format file so the next automation process can get it. I tried to use -machine-readable but not really get how to read the output easily.  Could anyone suggest a convenient way to do that? 

Thanks

Wilson

[Demetri Mouratis]

Ha. I came here to ask the same question.

Building AMIs with packer out of Jenkins. I need to have packer output the AMI ID to a file which I can then read (or some other work flow, just an idea). The next step for me is to run a aws cloudormation template passing in the ami. Then, update the resulting template to S3.

Any ideas appreciated.

-D

[Kyle Gochenour]

I was able to get the following to pull the AMI id from the packer output, assuming the output is dropped to a file. But I can not figure out how to get Jenkins to read the packer output from the build successfully.

awk -F"," '$3 ~/artifact/ && $5 ~/id/ && $6 ~/ami/ {print substr($6,length($6)-11,length($6))}' <outpuffilename>

[Dan Rigsby]

Im looking for the same.  It would be nice if there was an option on the AMI builder itself to specify the output file to write the AMI Id out to.  Until, then i will also be looking at scrapping the console output, but i really need to it also write out to our Bamboo server.

[Dan Rigsby]

This worked for me:

https://gist.github.com/danrigsby/11354917

It will write the output to stdout so it can be picked up my jenkins or bamboo or whatever.  It also writes to a output.txt file that used by the second command to write just the AMI ID to 'ami.txt'.

On Tuesday, March 25, 2014 3:50:51 AM UTC-4, Weng Wilson wrote:

[Alexander Gray II]

Thanks for finding a solution for this.  Much appreciated.  If packer ever changes the output, then these parsing commands will fail.  Too bad packer does not produce a json output (or something similar) to get this information.