microsoft / microsoft/vscode-remote-release
Dev Containers 0.459.1 / VS Code 1.123.0: attached-container reconnect still intermittently fails with "Reconnect attempt without old result"
@chrmarti is already working on this.
Since Jun 9, 2026.
- Dominant language
- Dockerfile
- Stars
- 4.2k
- Forks
- 469
- Avg merge
- 1d 1h
- Merged PRs (30d)
- 1
Description
Summary
This issue appears to still happen with VS Code 1.123.0 + Dev Containers 0.459.1, despite a maintainer comment in #11601 saying it appeared resolved with that version combination.
The important difference is that my failure is on the attached-container authority path, not the dev-container authority path from #11601.
In my case:
-
Reload With Extensions Disabledalmost never fails. -
Normal
Developer: Reload Windowalmost always fails. -
Starting from a fresh
Dev Containers: Attach to Running Containersometimes succeeds and sometimes fails. -
The first
resolveAuthority(attached-container,1)can succeed. -
After local extensions start activating, the extension host is terminated normally.
-
The second
resolveAuthority(attached-container,2)then fails with:no resolver for attached-containerReconnect attempt without old result
This looks like an attached-container reconnect race or lifecycle issue that is triggered or amplified by normal extension activation order.
Environment
Version: 1.123.0 (user setup)
Commit: 6a44c352bd24569c417e530095901b649960f9f8
Date: 2026-06-03T11:29:03+02:00
Electron: 42.2.0
ElectronBuildId: 14159160
Chromium: 148.0.7778.97
Node.js: 24.15.0
V8: 14.8.178.14-electron.0
OS: Windows_NT x64 10.0.19045
Dev Containers extension:
ms-vscode-remote.remote-containers@0.459.1
Remote mode:
Dev Containers: Attach to Running Container
Remote authority: attached-container
Docker context: non-default Docker context
Other Remote-related extensions installed:
ms-vscode-remote.remote-containers@0.459.1
ms-vscode-remote.remote-ssh@0.124.0
ms-vscode-remote.remote-ssh-edit@0.87.0
ms-vscode-remote.remote-wsl@0.104.3
ms-vscode-remote.vscode-remote-extensionpack@0.26.0
ms-vscode.remote-explorer@0.5.0
ms-vscode.remote-server@1.5.3
Some extensions that are active around the failure window in the logs:
formulahendry.auto-close-tag@0.5.15
formulahendry.auto-rename-tag@0.1.10
wayou.vscode-todo-highlight@1.0.5
vscodevim.vim@1.32.4
usernamehw.errorlens@3.28.0
PKief.material-icon-theme@5.35.0
ms-vscode-remote.remote-wsl@0.104.3
Relationship to #11601
#11601 was closed after a comment saying this appeared resolved with:
VS Code 1.123.0 + Dev Containers 0.459.1
However, I can still reproduce a very similar failure with the same VS Code and Dev Containers versions.
The difference is:
#11601: resolveAuthority(dev-container,2)
My case: resolveAuthority(attached-container,2)
So this may be a related but distinct bug in the attached-container reconnect path.
Observed behavior
Normal reload often fails to reconnect to an already running attached container.
The failure pattern is:
- VS Code activates
ms-vscode-remote.remote-containersforattached-container. PendingMigrationError: navigator is now a global in nodejsis logged from the Dev Containers extension.- First
resolveAuthority(attached-container,1)succeeds. - VS Code returns a local endpoint, for example
127.0.0.1:<port>. - Other local extensions start activating.
- The extension host receives a terminate message from renderer and exits with code 0.
- VS Code starts a new extension host.
- The second
resolveAuthority(attached-container,2)fails because there is initially no resolver forattached-container. - After
ms-vscode-remote.remote-containersactivates again, final resolve fails withReconnect attempt without old result.
Relevant log excerpt
2026-06-09 21:30:51.933 [info] ExtensionService#_doActivateExtension ms-vscode-remote.remote-containers, startup: false, activationEvent: 'onResolveRemoteAuthority:attached-container'
2026-06-09 21:30:52.098 [error] PendingMigrationError: navigator is now a global in nodejs, please see https://aka.ms/vscode-extensions/navigator for additional info on this error.
at get (.../extensionHostProcess.js:833:7243)
at it (...\.vscode\extensions\ms-vscode-remote.remote-containers-0.459.1\dist\extension\extension.js:20:5257)
...
2026-06-09 21:30:52.144 [info] [resolveAuthority(attached-container,1)][0ms] activating remote resolvers attached-container+...
2026-06-09 21:30:52.144 [info] [resolveAuthority(attached-container,1)][0ms] invoking final resolve()...
2026-06-09 21:30:57.939 [info] [resolveAuthority(attached-container,1)][5795ms] setting tunnel factory...
2026-06-09 21:30:57.939 [info] [resolveAuthority(attached-container,1)][5795ms] returned 127.0.0.1:40045
2026-06-09 21:31:00.115 [info] ExtensionService#_doActivateExtension formulahendry.auto-close-tag, startup: true, activationEvent: '*'
2026-06-09 21:31:00.117 [info] ExtensionService#_doActivateExtension formulahendry.auto-rename-tag, startup: true, activationEvent: '*'
2026-06-09 21:31:00.139 [info] ExtensionService#_doActivateExtension wayou.vscode-todo-highlight, startup: true, activationEvent: '*'
2026-06-09 21:31:00.473 [info] ExtensionService#_doActivateExtension ms-vscode-remote.remote-wsl, startup: false, activationEvent: 'onStartupFinished'
2026-06-09 21:31:00.581 [info] Extension host terminating: received terminate message from renderer
2026-06-09 21:31:00.581 [info] ExtensionService#_doActivateExtension usernamehw.errorlens, startup: false, activationEvent: 'onStartupFinished'
2026-06-09 21:31:00.581 [info] Extension host with pid 44316 exiting with code 0
2026-06-09 21:31:00.634 [info] Extension host terminating: received terminate message from renderer
2026-06-09 21:31:00.645 [info] Extension host with pid 27100 exiting with code 0
2026-06-09 21:31:01.157 [info] Extension host with pid 13588 started
2026-06-09 21:31:01.157 [info] [resolveAuthority(attached-container,2)][0ms] activating remote resolvers attached-container+...
2026-06-09 21:31:01.164 [error] [resolveAuthority(attached-container,2)][11ms] no resolver for attached-container undefined
2026-06-09 21:31:01.181 [info] Extension host with pid 12608 started
2026-06-09 21:31:01.209 [info] ExtensionService#_doActivateExtension ms-vscode-remote.remote-containers, startup: false, activationEvent: 'onResolveRemoteAuthority:attached-container'
2026-06-09 21:31:01.404 [info] [resolveAuthority(attached-container,2)][226ms] invoking final resolve()...
2026-06-09 21:31:01.404 [error] [resolveAuthority(attached-container,2)][226ms] returned an error Reconnect attempt without old result.
Steps to reproduce
I do not yet have a fully minimized extension set, but the issue is consistently correlated with normal extension activation.
Current observed repro:
- Open VS Code 1.123.0 with Dev Containers 0.459.1.
- Use
Dev Containers: Attach to Running Container. - Attach to an already running container using the
attached-containerauthority. - Wait until the window is fully loaded.
- Run
Developer: Reload Window. - Observe that reconnect often fails with
Cannot reconnect/Reconnect attempt without old result.
Control case:
- Run
Developer: Reload With Extensions Disabled. - Attach/reload the same running container.
- Reconnect almost never fails.
Additional observation:
- Directly attaching without reload is intermittent.
- Normal reload with extensions enabled almost always fails.
- Reload with extensions disabled is much more stable.
Expected behavior
Reloading the VS Code window should reconnect to the already running attached container.
If the old resolve result is unavailable after extension host restart, the attached-container resolver should either recover by performing a fresh attach or fail with a more actionable diagnostic. It should not get permanently stuck in Reconnect attempt without old result while the container is still running.
Actual behavior
VS Code fails to reconnect even though the container is still running.
The second resolveAuthority(attached-container,2) fails with:
no resolver for attached-container
Reconnect attempt without old result.
Why I think this is still a bug in 1.123.0 + 0.459.1
I understand #11601 was reported as resolved with VS Code 1.123.0 + Dev Containers 0.459.1. However, this issue still reproduces in that exact version combination in the attached-container path.
The fact that Reload With Extensions Disabled almost never fails suggests that normal extension activation timing can trigger or amplify the race. But the failure still occurs in the Dev Containers remote authority resolution path, and the attached container remains running.
The resolver appears to lose or require an old resolve result across extension host restart, then cannot recover.
Request
Could you please check whether the fix/verification for #11601 covered only the dev-container authority path?
It would be useful to know whether attached-container reconnect is expected to support recovery after local extension host restart, especially when the first resolve already succeeded and returned a local endpoint.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.