Multiple containers in none state at worker nodes


#1
## Version Info
* Concourse version: 3.8
* Deployment type (BOSH/Docker/binary): binary
* Infrastructure/IaaS: hosted on vms (2 x Web and 4 x Worker)

I see multiple containers running on the worker nodes (most of them are in none state) and was wondering why it’s not getting terminated. I have only a couple of pipelines running currently and I am sure that those would not be spinning up these many containers. I would appreciate if someone could point out what’s going wrong here.

./fly -t new  containers
  handle                              worker     pipeline  job    build #  build id  type   name   attempt
00295365-5fca-4b0a-5ef2-f595aff4d905  prd-conw2	 none      none   none     none      check  none   n/a
0056dc87-ba36-40d4-6b05-e58010d9a3b1  prd-conw4	 none      none   none     none      check  none   n/a
00d54d9c-9b6c-4b42-7db0-b24dccf518d2  prd-conw1	 none      none   none     none      check  none   n/a
011bc07b-5801-488b-6eda-bca19aff1843  prd-conw1	 none      none   none     none      check  none   n/a
0199cbd4-aee8-42da-46ea-0fd78672ac4c  prd-conw2	 none      none   none     none      check  none   n/a
01b14097-a15a-45b4-43cc-1dc6933c33a5  prd-conw3	 none      none   none     none      check  none   n/a
01b607fe-c848-4a8b-6ec4-dfdbd163210c  prd-conw3	 none      none   none     none      check  none   n/a
01b817d3-98fb-4258-4022-fdb4e85d4172  prd-conw3	 none      none   none     none      get    none   n/a
01b87cda-2562-4778-79d5-fafb63d47fd2  prd-conw4	 none      none   none     none      check  none   n/a
01d5ebb2-14eb-46b4-769c-75fe18665a0d  prd-conw4	 none      none   none     none      check  none   n/a
01edd08a-25b9-482b-5e3d-dc23eb1818f2  prd-conw2	 none      none   none     none      check  none   n/a
0206004d-af32-4c7f-66ad-04d892dbfb43  prd-conw1	 none      none   none     none      check  none   n/a
0239c916-2355-401f-6e97-03486ff5953b  prd-conw1	 none      none   none     none      get    none   n/a
025cdbde-ccb6-4647-5826-a76cd1aea34d  prd-conw4	 none      none   none     none      get    none   n/a
02bcf4fb-dfe6-4796-7c8b-475345728c07  prd-conw1	 none      none   none     none      check  none   n/a

#2

I don’t think there is anything wrong. Most of the containers listed are check containers for which you would expect there to be one per resource config. The reason they aren’t getting terminated is that resources check every minute by default.


#3

Yep. We’re planning on making this less confusing with https://github.com/concourse/concourse/issues/3079


#4

I could see that we have around 1K containers in total and almost all are in check or none. Is this normal ?