Cli: Docker login to registry in windows server 2016 - Error saving credentials

[RotemEmergi]

Hi ,

I am using windows server 2016 with docker for windows .

docker info
Containers: 2
Running: 0
Paused: 0
Stopped: 2
Images: 105
Server Version: 18.06.0-ce
Storage Driver: overlay2
Backing Filesystem: extfs
Supports d_type: true
Native Overlay Diff: true
Logging Driver: json-file
Cgroup Driver: cgroupfs
Plugins:
Volume: local
Network: bridge host macvlan null overlay
Log: awslogs fluentd gcplogs gelf journald json-file logentries splunk syslog
Swarm: inactive
Runtimes: runc
Default Runtime: runc
Init Binary: docker-init
containerd version: d64c661f1d51c48782c9cec8fda7604785f93587
runc version: 69663f0bd4b60df09991c08812a60108003fa340
init version: fec3683
Security Options:
seccomp
Profile: default
Kernel Version: 4.9.93-linuxkit-aufs
Operating System: Docker for Windows
OSType: linux
Architecture: x86_64
CPUs: 2
Total Memory: 1.934GiB
Name: linuxkit-00155d000400
ID: LDIE:JDCD:KNOR:RNPH:7MWV:MNKE:I2LH:3NMM:ZERH:M7HE:MWAI:Q4RW
Docker Root Dir: /var/lib/docker
Debug Mode (client): false
Debug Mode (server): true
File Descriptors: 23
Goroutines: 48
System Time: 2018-08-02T14:33:25.5407319Z
EventsListeners: 1
Registry: https://index.docker.io/v1/
Labels:
Experimental: false
Insecure Registries:
127.0.0.0/8
Live Restore Enabled: false

I am using jenkins pipeline for build docker images and it's build and tag the images but when i am trying to do docker login before push, i am getting this error :

"WARNING! Using --password via the CLI is insecure. Use --password-stdin.
Error saving credentials: error storing credentials - err: exit status 1, out: A specified logon session does not exist. It may already have been terminated."

This is not happen to me when i am with logon active user .

Thanks ,

[RotemEmergi]

Any update ?

[RotemEmergi]

Any update ?

[deevodavis71]

I'm getting the same error... can't find a response. Help!

[RotemEmergi]

Any update ?

[mrysanek]

Jumping on the bandwagon here... same issue, using docker plugin and jenkinsfile:

stage('Dockerfile Push') {
pipelineParams['registries'].each { reg ->
echo "Pushing to registry: ${reg['name']}"
docker.withRegistry("https://${reg['name']}", reg['credentialsId']) {

Error happens on the last line of above:

$ docker login -u * -p * https://dockerinternal.sanitized.com
WARNING! Using --password via the CLI is insecure. Use --password-stdin.
Error saving credentials: error storing credentials - err: exit status 1, out: A specified logon session does not exist. It may already have been terminated.

[mrysanek]

Followup: This only happens through an ssh session to the Windows server. When the same command is run locally, the login succeeds. When an ssh session is established directly to the server (NOT using Jenkins) the same error is observed.

[mrysanek]

Digging into the server's event log I got this:

System.ComponentModel.Win32Exception (0x80004005): Failed to write credentials
at Microsoft.Alm.Authentication.BaseSecureStore.WriteCredential(String targetName, Credential credentials)
at Microsoft.Alm.Authentication.SecretStore.WriteCredentials(TargetUri targetUri, Credential credentials)
at Microsoft.Alm.Cli.Program.Store()
at Microsoft.Alm.Cli.Program.Run(String[] args)

Log Name: Application
Source: Git Credential Manager
Date: 9/17/2018 10:49:33 AM
Event ID: 0
Task Category: None
Level: Error
Keywords: Classic
User: N/A
Computer: **.com

I am using OpenSSH running on the server, and found a similar problem (albeit with a git push for client rather than docker login):

https://github.com/Microsoft/Git-Credential-Manager-for-Windows/issues/502

My issue is, in fact, the user and owner are different when I run this from a powershell launched through the ssh session: [Security.Principal.WindowsIdentity]::GetCurrent()
This is the reason Jenkins pipeline docker usage fails for me (beause it's connected through an ssh session).
When the above is run in a powershell spawned through an RDP session, the user and owner are the same.

@RotemEmergi @deevodavis71 Either of you using same access method to the Windows server (OpenSSH)?

[RotemEmergi]

I am using with ssh

[DankoValkov]

Hey Guys,

I have had the same issue, when trying to authenticate from a PowerShell session, while it was working fine when the same command was executed locally.

The way I have fixed it, is by looking at the C:\Users_youruser_.docker\config.json
If there is a line like: "credsStore": "wincred" , make sure that you change it to the following:
"credsStore": "", save the config, and re-execute the same command locally:
$ docker login -u * -p * url

Afterwards take a look at the config again, and now you should have the following line instead, under the repository:
"auth": "...."

Having the config that way, allowed me to execute successfully the commands from another session as well.

[samjgalbraith]

I have the same problem, except it doesn't seem to have anything to do with SSH. Docker CLI will sometimes just start doing giving me this error when trying to log into an AWS ECR repository that has worked previously. The only thing I can find that makes it come right again (for a while) is to reinstall Docker. It's really annoying. I tried deleting the credentials that Docker stores in Windows Credentials Manager but that doesn't seem to help at all.

[DankoValkov]

Changing the credentials that are stored in "Windows Credentials Manager" wasn't helpful for my case as well, that's why I had to change the config file of the Docker directly on that environment.

After making sure that the password written in the login command matches the one inside the config, under the "auth": "yy==" like:
docker login -u ** -p yy== url
resolved the issue for me, and I was able to both execute the command successfully on prem, and remotely as well.

[basilin]

Having the issue while executing via ssh.

[pleonex]

EDIT: The workaround doesn't work anymore. Docker always overwrites now the credsStore when it's empty.

The workaround of setting credsStore to "" works however in the latest version of Docker (2.2.0) when you restart Docker it will reset that field to desktop failing again...

For my case this workaround is not useful since every time we reboot the build machine we will need to log in and manually change that value or write some kind of script and put it everywhere in our pipeline.

My workaround for the workaround is to create a separate .config.json file with the property empty in a non-default location and run all the Docker commands pointing to that config file so it will use it but it won't find it to reset the value on restart.

[owlyowl]

Hi all I have this exact issue on Windows 10 I:

1. Can see the credential in the windows credential manager
2. Have tried clearing that credential and logging back in again
3. Have checked windows event viewer with nothing of note
4. Have tried clearing out docker.config's entry for credsStore
5. Have checked that Network access: Do not allow storage of passwords and credentials for network authentication is disabled

Has anyone found a resolution for this issue it is driving me crazy

[jredp]

Same exact thing for me. Tried all the fixes.

FYI I am using TeamCity to fire off a Powershell Script on that server and it fails the build, but when I go to the server and manually run it, it works fine. It's also sporadic. Works fine some builds, then magically one day it doesn't work.