fix(code): probe unreachable Ollama daemon once per reload - #4806
Merged
Mason Daugherty (mdrxy) merged 4 commits intoJul 17, 2026
Conversation
Ollama discovery runs from two startup callers (`get_available_models` and `get_model_profiles`), each hitting `_get_ollama_installed_models`. Because empty results were never cached, an unreachable daemon was probed twice per reload and logged "not detected" twice, interleaving confusingly with the "returned no models" line from the caller. Track endpoints that fail the TCP presence preflight and cache their empty result so the probe and log happen once per reload. A reachable daemon reporting no models is still left uncached so a later `ollama pull` is discovered without `/reload`. Co-authored-by: open-swe[bot] <open-swe@users.noreply.github.com>
github-actions
Bot
added
dcode
Mason Daugherty (mdrxy)
marked this pull request as ready for review
July 16, 2026 21:47
Mason Daugherty (mdrxy)
deleted the
mdrxy/code/ollama-discovery-negative-cache
branch
July 17, 2026 18:39
Mason Daugherty (mdrxy)
pushed a commit
that referenced
this pull request
Jul 17, 2026
> [!CAUTION] > Merging this PR will automatically publish to **PyPI** and create a **GitHub release**. For the full release process, see [`.github/RELEASING.md`](https://github.com/langchain-ai/deepagents/blob/main/.github/RELEASING.md). --- _Release notes preview: keep this section in sync with the package `CHANGELOG.md`. The published GitHub release body is extracted from the merged `CHANGELOG.md` by `release.yml`, not from this PR description._ --- ## [0.1.43](deepagents-code==0.1.42...deepagents-code==0.1.43) (2026-07-17) ### Features - Added classifier-backed Auto approval mode behind `DEEPAGENTS_CODE_EXPERIMENTAL=1` ([#4804](#4804)). - Added a shutdown toast for deferred exits ([#4830](#4830)). - Task descriptions that were truncated can now be expanded by clicking or pressing `Ctrl+O` ([#4811](#4811)). - Debug Console clears with `Ctrl+L` now persist after reopening ([#4812](#4812)). - Added debug logging for skill-name override collisions ([#4772](#4772)). ### Bug Fixes - Keep chat input responsive during `/restart` ([#4808](#4808)). - Fixed paste placeholders disappearing when backspacing a newline below them ([#4757](#4757)). - Made markdown `AppMessage` output selectable ([#4814](#4814)). - Fixed live tool-group counts to include only running tools ([#4809](#4809)). - Kept `task` timers monotonic across nested subagent human-in-the-loop flows ([#4771](#4771)). - Preserved goal criteria proposals when marker clearing fails ([#4785](#4785)). - Reduced repeated probing of an unreachable Ollama daemon to once per reload ([#4806](#4806)). - Quieted MCP auth-skip debug logging for known patterns ([#4805](#4805)). - Improved `/version` diagnostics for editable installs and core dependency reporting, including surfacing `langchain-quickjs` ([#4816](#4816), [#4813](#4813)). - Removed the `uv install` tip from the `/version` update hint ([#4822](#4822)). _End release notes preview._ --- > [!NOTE] > A **New Contributors** section is appended to the GitHub release notes automatically at publish time (see [Release Pipeline](https://github.com/langchain-ai/deepagents/blob/main/.github/RELEASING.md#release-pipeline), step 2). --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: langchain-oss-automated-triage[bot] <248757908+langchain-oss-automated-triage[bot]@users.noreply.github.com>
Marcelo5444
pushed a commit
to Marcelo5444/deepagents
that referenced
this pull request
Jul 30, 2026
…-ai#4806) When an Ollama daemon is down, discovery no longer probes it twice or emits duplicate, out-of-order debug logs at startup — it probes and logs "not detected" once per reload. --- Ollama discovery runs from two startup callers — `get_available_models` and `get_model_profiles` — that both call `_get_ollama_installed_models`. Empty results were never cached, so an unreachable daemon was probed twice per reload and logged `Ollama daemon not detected ...` twice. Because the second `not detected` (from `get_model_profiles`) landed right after the first pass's `discovery returned no models` line, the two appeared out of order even though each individual pass logs in the correct order. This tracks endpoints that fail the TCP presence preflight in `_ollama_unreachable_endpoints` and negatively-caches their empty result, so the probe and log happen once. A *reachable* daemon that simply reports no models is still left uncached, preserving recovery of a later `ollama pull` without `/reload` (covered by the existing `test_empty_installed_model_discovery_not_cached`). Made by [Open SWE](https://openswe.vercel.app/agents/7b64e449-2a7e-f258-17a1-5ffe933608a3) --------- Co-authored-by: open-swe[bot] <open-swe@users.noreply.github.com>
Marcelo5444
pushed a commit
to Marcelo5444/deepagents
that referenced
this pull request
Jul 30, 2026
> [!CAUTION] > Merging this PR will automatically publish to **PyPI** and create a **GitHub release**. For the full release process, see [`.github/RELEASING.md`](https://github.com/langchain-ai/deepagents/blob/main/.github/RELEASING.md). --- _Release notes preview: keep this section in sync with the package `CHANGELOG.md`. The published GitHub release body is extracted from the merged `CHANGELOG.md` by `release.yml`, not from this PR description._ --- ## [0.1.43](langchain-ai/deepagents@deepagents-code==0.1.42...deepagents-code==0.1.43) (2026-07-17) ### Features - Added classifier-backed Auto approval mode behind `DEEPAGENTS_CODE_EXPERIMENTAL=1` ([langchain-ai#4804](langchain-ai#4804)). - Added a shutdown toast for deferred exits ([langchain-ai#4830](langchain-ai#4830)). - Task descriptions that were truncated can now be expanded by clicking or pressing `Ctrl+O` ([langchain-ai#4811](langchain-ai#4811)). - Debug Console clears with `Ctrl+L` now persist after reopening ([langchain-ai#4812](langchain-ai#4812)). - Added debug logging for skill-name override collisions ([langchain-ai#4772](langchain-ai#4772)). ### Bug Fixes - Keep chat input responsive during `/restart` ([langchain-ai#4808](langchain-ai#4808)). - Fixed paste placeholders disappearing when backspacing a newline below them ([langchain-ai#4757](langchain-ai#4757)). - Made markdown `AppMessage` output selectable ([langchain-ai#4814](langchain-ai#4814)). - Fixed live tool-group counts to include only running tools ([langchain-ai#4809](langchain-ai#4809)). - Kept `task` timers monotonic across nested subagent human-in-the-loop flows ([langchain-ai#4771](langchain-ai#4771)). - Preserved goal criteria proposals when marker clearing fails ([langchain-ai#4785](langchain-ai#4785)). - Reduced repeated probing of an unreachable Ollama daemon to once per reload ([langchain-ai#4806](langchain-ai#4806)). - Quieted MCP auth-skip debug logging for known patterns ([langchain-ai#4805](langchain-ai#4805)). - Improved `/version` diagnostics for editable installs and core dependency reporting, including surfacing `langchain-quickjs` ([langchain-ai#4816](langchain-ai#4816), [langchain-ai#4813](langchain-ai#4813)). - Removed the `uv install` tip from the `/version` update hint ([langchain-ai#4822](langchain-ai#4822)). _End release notes preview._ --- > [!NOTE] > A **New Contributors** section is appended to the GitHub release notes automatically at publish time (see [Release Pipeline](https://github.com/langchain-ai/deepagents/blob/main/.github/RELEASING.md#release-pipeline), step 2). --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: langchain-oss-automated-triage[bot] <248757908+langchain-oss-automated-triage[bot]@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
When an Ollama daemon is down, discovery no longer probes it twice or emits duplicate, out-of-order debug logs at startup — it probes and logs "not detected" once per reload.
Ollama discovery runs from two startup callers —
get_available_modelsandget_model_profiles— that both call_get_ollama_installed_models. Empty results were never cached, so an unreachable daemon was probed twice per reload and loggedOllama daemon not detected ...twice. Because the secondnot detected(fromget_model_profiles) landed right after the first pass'sdiscovery returned no modelsline, the two appeared out of order even though each individual pass logs in the correct order.This tracks endpoints that fail the TCP presence preflight in
_ollama_unreachable_endpointsand negatively-caches their empty result, so the probe and log happen once. A reachable daemon that simply reports no models is still left uncached, preserving recovery of a laterollama pullwithout/reload(covered by the existingtest_empty_installed_model_discovery_not_cached).Made by Open SWE