[
https://issues.apache.org/jira/browse/KUDU-3806?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18137971#comment-18137971
]
ASF subversion and git services commented on KUDU-3806:
-------------------------------------------------------
Commit 0bf267597529dd3ea783f12bc36adcf4c50b5372 in kudu's branch
refs/heads/master from Marton Greber
[ https://gitbox.apache.org/repos/asf?p=kudu.git;h=0bf267597 ]
KUDU-3806: split CLI actions into a library
Move the top-level mode list out of tool_main.cc's RootMode() into a
new BuildRootMode() (tool_action_root.cc), and build the CLI action
factories into a kudu_cli_actions library, with the 'kudu' executable
reduced to a thin main() linked against it.
This lets unit tests reconstruct the full action tree in-process by
linking the library, instead of recompiling every tool_action_*.cc
source into each test. It is also the single source of truth for the
top-level mode list, so a newly added mode is wired in one place
rather than mirrored across main() and its consumers.
Pure refactor: the 'kudu' binary is unchanged. tool_action_root.cc
lives in its own translation unit (part of the CLI sources) rather
than in the low-level kudu_tools_util library, so referencing every
Build*Mode() factory does not inject undefined symbols into
kudu_tools_util consumers such as ksck.
Change-Id: I3c47cf1844042fc09493edf18652493742187f03
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Reviewed-on: https://gerrit.cloudera.org/24906
Reviewed-by: Zoltan Chovan <[email protected]>
Tested-by: Marton Greber <[email protected]>
Reviewed-by: Zoltan Martonka <[email protected]>
> Kudu MCP
> --------
>
> Key: KUDU-3806
> URL: https://issues.apache.org/jira/browse/KUDU-3806
> Project: Kudu
> Issue Type: New Feature
> Reporter: Marton Greber
> Assignee: Marton Greber
> Priority: Major
>
> Introduce kudu mcp serve, an MCP server that exposes existing kudu CLI
> actions as MCP tools so operators can drive cluster diagnostics and
> administration from an MCP-capable client (e.g. an LLM agent) without
> hand-crafting command lines.
> Design highlights:
> - Safety classification. Every action in the CLI action tree resolves to
> exactly one Disposition:
> -- SURFACE — read-only, always exposed (node-local reads are surfaced but
> tagged as only meaningful on that node);
> -- GATED — cluster-mutating, hidden unless --allow-writes is set;
> -- REJECT — long-running/never-promptly-returning ops that would wedge the
> single-threaded serve loop;
> -- EXCLUDE — interactive or node-local mutating actions, never surfaced.
> - Read-only by default. Mutating tools are only advertised with
> --allow-writes.
> - Process isolation. Each tools/call runs its action in a fresh child kudu
> process bounded by --mcp_tool_timeout_sec (default 60s), so a hung action
> cannot freeze the server.
> - Credential hygiene. --master_addresses is registered once at serve time
> and injected into each call, so it never has to appear in the conversation.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)