From ec3df8e22acc72ec14c4ae966e43d9c64923a42e Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 13 Jul 2026 09:13:03 +0000 Subject: [PATCH 1/5] prose: document `ant node dismiss` and node status health/eviction output ant-client added the `ant node dismiss ` command and folded a fleet-health summary plus node-eviction reporting into `ant node status`. Document both on the CLI command reference and refresh its verification stamp to the audited head SHA. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_018k8QXTGkcfKXvycW1CAwRn --- docs/cli/command-reference.md | 29 ++++++++++++++++++++++++++--- 1 file changed, 26 insertions(+), 3 deletions(-) diff --git a/docs/cli/command-reference.md b/docs/cli/command-reference.md index 930d46b..fd0ca7b 100644 --- a/docs/cli/command-reference.md +++ b/docs/cli/command-reference.md @@ -3,8 +3,8 @@ @@ -21,6 +21,7 @@ ant │ │ ├── stop │ │ ├── status │ │ └── info +│ ├── dismiss │ ├── reset │ ├── start │ ├── status @@ -336,7 +337,7 @@ ant node start --service-name node1 ### `ant node status` -Shows the status of all registered nodes. +Shows the status of all registered nodes. Each node reports its state, which includes `Running`, `Stopped`, `Starting`, `Stopping`, `Errored`, and `Evicted`. A node is evicted when its host runs low on disk; the row shows the eviction reason and the exact command to clear it. When the daemon is running, the output also opens with a fleet-health summary of `Healthy`, `Warning`, or `Critical`, followed by a line for each check that is not healthy. **Parameters:** @@ -348,6 +349,28 @@ This command has no command-specific parameters. ant node status ``` +Add `--json` (a root flag, before the subcommand) to get a machine-readable payload. The JSON object carries the per-node list under `nodes`, the `total_running` and `total_stopped` counts, and a `health` object with the fleet-health summary (or `null` when the daemon is not running). + +```bash +ant --json node status +``` + +### `ant node dismiss` + +Removes an evicted node from the registry so it no longer appears in `ant node status`. Run this after a node is evicted for low disk and you have reclaimed space or no longer want the node. When the daemon is running, the command clears the node from the daemon's in-memory registry as well; when the daemon is stopped, it operates directly on the registry file. + +**Parameters:** + +| Name | Type | Required | Description | +|------|------|----------|-------------| +| `` | integer | Yes | The ID of the evicted node to dismiss. This is the numeric ID shown in the `ant node status` list. | + +**Example:** + +```bash +ant node dismiss 3 +``` + ### `ant node stop` Stops all registered nodes, or one named node with `--service-name`. From e88dc2c278ca83300b2bcc0a72c1188434ffec74 Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 13 Jul 2026 10:18:01 +0000 Subject: [PATCH 2/5] prose: correct node status JSON contract, dismiss semantics, and heading Address review of the CLI command reference: - Qualify `health` null and the terminal fleet-health summary: null also occurs when the daemon is running but the snapshot cannot be retrieved. - Correct `ant node dismiss` semantics: not evicted-only. Daemon running removes any non-running node (and refuses a running one); daemon stopped removes the registry entry directly and does not stop a running process. - Add a full JSON schema and a valid example for `ant --json node status`, with actual lowercase wire values for statuses and health levels and verified total_running/total_stopped count semantics. - Explain automatic eviction (stops node, deletes data dir, retains the registry record) and the dismiss-then-add recovery path. - Replace the inline-code dismiss heading with an outcome-oriented heading plus the literal `ant node dismiss ` syntax, matching the reference-page convention. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_018k8QXTGkcfKXvycW1CAwRn --- docs/cli/command-reference.md | 112 ++++++++++++++++++++++++++++++++-- 1 file changed, 107 insertions(+), 5 deletions(-) diff --git a/docs/cli/command-reference.md b/docs/cli/command-reference.md index fd0ca7b..e1fa8ff 100644 --- a/docs/cli/command-reference.md +++ b/docs/cli/command-reference.md @@ -337,7 +337,9 @@ ant node start --service-name node1 ### `ant node status` -Shows the status of all registered nodes. Each node reports its state, which includes `Running`, `Stopped`, `Starting`, `Stopping`, `Errored`, and `Evicted`. A node is evicted when its host runs low on disk; the row shows the eviction reason and the exact command to clear it. When the daemon is running, the output also opens with a fleet-health summary of `Healthy`, `Warning`, or `Critical`, followed by a line for each check that is not healthy. +Shows the status of all registered nodes. Each node reports a state, shown in the table as `Running`, `Stopped`, `Starting`, `Stopping`, `Errored`, or `Evicted`. An evicted node is one the daemon automatically stopped when its host ran low on disk: the daemon deletes that node's data directory to reclaim space and keeps its registry record marked `Evicted`. The row for an evicted node shows the eviction reason and the exact `ant node dismiss` command to clear it. + +When the daemon is running and a health snapshot is available, the output opens with a fleet-health summary of `Healthy`, `Warning`, or `Critical`, followed by one line per check that is not healthy. The summary is omitted when the daemon is stopped, and also when the daemon is running but the snapshot cannot be retrieved. **Parameters:** @@ -349,21 +351,121 @@ This command has no command-specific parameters. ant node status ``` -Add `--json` (a root flag, before the subcommand) to get a machine-readable payload. The JSON object carries the per-node list under `nodes`, the `total_running` and `total_stopped` counts, and a `health` object with the fleet-health summary (or `null` when the daemon is not running). +**JSON output:** + +Add `--json` (a root flag, so it comes before the subcommand) to emit a machine-readable payload. ```bash ant --json node status ``` -### `ant node dismiss` +The payload has these top-level fields: + +| Field | Type | Description | +|------|------|-------------| +| `nodes` | array | One object per registered node. | +| `total_running` | integer | Count of nodes reported as `running` or `starting`. | +| `total_stopped` | integer | Count of every other node, including `stopped`, `stopping`, `errored`, and `evicted`. | +| `health` | object or null | Fleet-health snapshot. It is `null` when the daemon is stopped, and also when the daemon is running but the snapshot cannot be retrieved, so `null` alone does not mean the daemon is stopped. | + +Each `nodes` entry has these fields: + +| Field | Type | Description | +|------|------|-------------| +| `node_id` | integer | Node ID used by the other `ant node` commands. | +| `name` | string | Service name of the node. | +| `version` | string | Node binary version. | +| `status` | string | One of `running`, `stopped`, `starting`, `stopping`, `errored`, `evicted`. | +| `pid` | integer | Process ID. Present only while the node is running. | +| `uptime_secs` | integer | Seconds since the node process started. Present only while the node is running. | +| `eviction` | object | Eviction detail. Present only when `status` is `evicted`. | + +An `eviction` object has these fields: + +| Field | Type | Description | +|------|------|-------------| +| `reason` | string | Human-readable explanation of the eviction. | +| `evicted_at` | integer | Unix epoch seconds at which the eviction occurred. | +| `reclaimed_bytes` | integer | Approximate bytes reclaimed by deleting the node's data directory. | + +When present, the `health` object has an `overall` level (`green`, `warning`, or `critical` — the worst level across all checks) and a `checks` array. Each `checks` entry has these fields: + +| Field | Type | Description | +|------|------|-------------| +| `kind` | string | Check type. Currently `disk_space`. | +| `level` | string | `green`, `warning`, or `critical`. | +| `summary` | string | Human-readable, user-facing one-liner. | +| `partition` | string | Partition the finding concerns. Disk checks only. | +| `available_bytes` | integer | Free bytes on the partition. Disk checks only. | +| `eviction_threshold_bytes` | integer | Free-space floor at which an eviction triggers. Disk checks only. | +| `candidate` | object | The node that would be evicted next, when one applies. Has `node_id` (integer), `data_dir` (string), and `size_bytes` (integer, the space its eviction would free). | + +Example payload with one running node and one evicted node: + +```json +{ + "nodes": [ + { + "node_id": 1, + "name": "antnode1", + "version": "0.4.0", + "status": "running", + "pid": 48213, + "uptime_secs": 3600 + }, + { + "node_id": 2, + "name": "antnode2", + "version": "0.4.0", + "status": "evicted", + "eviction": { + "reason": "Low disk: 480 MiB free, evicted to reclaim space", + "evicted_at": 1720800000, + "reclaimed_bytes": 2147483648 + } + } + ], + "total_running": 1, + "total_stopped": 1, + "health": { + "overall": "warning", + "checks": [ + { + "kind": "disk_space", + "level": "warning", + "summary": "900 MiB free on /; node 1 is the next eviction candidate", + "partition": "/", + "available_bytes": 943718400, + "eviction_threshold_bytes": 524288000, + "candidate": { + "node_id": 1, + "data_dir": "/home/alice/.local/share/autonomi/node/antnode1", + "size_bytes": 1073741824 + } + } + ] + } +} +``` + +### Dismiss a node from the registry + +**Command:** `ant node dismiss ` + +Removes a node's registry entry so it no longer appears in `ant node status`. This is the recovery step for a node the daemon evicted for low disk, though it is not restricted to evicted nodes. + +Dismissal behaviour depends on the daemon: + +- With the daemon running, dismiss removes any node that is not currently running. It refuses to dismiss a running node and asks you to stop it first. +- With the daemon stopped, dismiss removes the registry entry directly and does not stop any process. Dismiss a node only when it is not running, so you do not leave an orphaned node process behind. -Removes an evicted node from the registry so it no longer appears in `ant node status`. Run this after a node is evicted for low disk and you have reclaimed space or no longer want the node. When the daemon is running, the command clears the node from the daemon's in-memory registry as well; when the daemon is stopped, it operates directly on the registry file. +To recover capacity after an eviction, dismiss the retained `Evicted` record, then add a replacement node with `ant node add` if you still need it. **Parameters:** | Name | Type | Required | Description | |------|------|----------|-------------| -| `` | integer | Yes | The ID of the evicted node to dismiss. This is the numeric ID shown in the `ant node status` list. | +| `` | integer | Yes | ID of the node to dismiss, as shown in the `ant node status` list. | **Example:** From 3325426ccd6d3a64a4c8e58662633cbd642ad201 Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 13 Jul 2026 11:46:48 +0000 Subject: [PATCH 3/5] prose: correct node status JSON example state model and eviction recovery Address remaining review of the CLI command reference: - Rebuild the `ant --json node status` example to a source-consistent state: two running nodes sharing a partition plus one evicted node, so the disk-space check can legitimately name a candidate (source only emits a candidate when at least two running nodes share a partition and the level is not green). Totals, health, candidate, and the summary/reason strings now match the health and eviction `format!` templates in ant-core. - Correct eviction and recovery guidance: eviction attempts cleanup; deletion can fail, leaving an Evicted record with reclaimed_bytes 0 and a manual-cleanup reason. Dismissing removes only the registry entry and does not reclaim space; free space manually when cleanup did not, and add a replacement node only after capacity is available. - Mark candidate/partition/available_bytes/eviction_threshold_bytes as optional and omitted when not applicable; describe the supported check type directly without time-relative wording. - behaviour -> behavior; "currently running" -> "running". - Add audited expected output for `ant node dismiss 3`. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_018k8QXTGkcfKXvycW1CAwRn --- docs/cli/command-reference.md | 52 ++++++++++++++++++++++------------- 1 file changed, 33 insertions(+), 19 deletions(-) diff --git a/docs/cli/command-reference.md b/docs/cli/command-reference.md index e1fa8ff..9272058 100644 --- a/docs/cli/command-reference.md +++ b/docs/cli/command-reference.md @@ -337,7 +337,7 @@ ant node start --service-name node1 ### `ant node status` -Shows the status of all registered nodes. Each node reports a state, shown in the table as `Running`, `Stopped`, `Starting`, `Stopping`, `Errored`, or `Evicted`. An evicted node is one the daemon automatically stopped when its host ran low on disk: the daemon deletes that node's data directory to reclaim space and keeps its registry record marked `Evicted`. The row for an evicted node shows the eviction reason and the exact `ant node dismiss` command to clear it. +Shows the status of all registered nodes. Each node reports a state, shown in the table as `Running`, `Stopped`, `Starting`, `Stopping`, `Errored`, or `Evicted`. An evicted node is one the daemon automatically stopped when its host ran low on disk. Eviction stops the node and attempts to delete its data directory to reclaim space, then keeps the node's registry record marked `Evicted`. The table row for an evicted node shows the eviction reason and the exact `ant node dismiss` command to clear it. Deletion can fail; check the eviction reason, and `reclaimed_bytes` in the JSON output, to confirm whether space was actually recovered. When the daemon is running and a health snapshot is available, the output opens with a fleet-health summary of `Healthy`, `Warning`, or `Critical`, followed by one line per check that is not healthy. The summary is omitted when the daemon is stopped, and also when the daemon is running but the snapshot cannot be retrieved. @@ -386,21 +386,21 @@ An `eviction` object has these fields: |------|------|-------------| | `reason` | string | Human-readable explanation of the eviction. | | `evicted_at` | integer | Unix epoch seconds at which the eviction occurred. | -| `reclaimed_bytes` | integer | Approximate bytes reclaimed by deleting the node's data directory. | +| `reclaimed_bytes` | integer | Approximate bytes reclaimed by deleting the node's data directory. `0` when the deletion did not succeed, in which case the reason explains that manual cleanup may be needed. | When present, the `health` object has an `overall` level (`green`, `warning`, or `critical` — the worst level across all checks) and a `checks` array. Each `checks` entry has these fields: | Field | Type | Description | |------|------|-------------| -| `kind` | string | Check type. Currently `disk_space`. | +| `kind` | string | Check type. The supported value is `disk_space`. | | `level` | string | `green`, `warning`, or `critical`. | | `summary` | string | Human-readable, user-facing one-liner. | -| `partition` | string | Partition the finding concerns. Disk checks only. | -| `available_bytes` | integer | Free bytes on the partition. Disk checks only. | -| `eviction_threshold_bytes` | integer | Free-space floor at which an eviction triggers. Disk checks only. | -| `candidate` | object | The node that would be evicted next, when one applies. Has `node_id` (integer), `data_dir` (string), and `size_bytes` (integer, the space its eviction would free). | +| `partition` | string | Optional. Opaque identifier of the partition the finding concerns. Present for `disk_space` checks; omitted for other check kinds. | +| `available_bytes` | integer | Optional. Free bytes on the partition. Present for `disk_space` checks; omitted for other check kinds. | +| `eviction_threshold_bytes` | integer | Optional. Free-space floor at which an eviction triggers. Present for `disk_space` checks; omitted for other check kinds. | +| `candidate` | object | Optional. Names the node that would be evicted next, with `node_id` (integer), `data_dir` (string), and `size_bytes` (integer, the space its eviction would free). Omitted when no candidate applies — a candidate is only produced when at least two running nodes share the partition and the level is not `green`. | -Example payload with one running node and one evicted node: +Example payload with two running nodes that share a partition and one previously evicted node. The disk-space check is at `warning`, so it names the smaller running node as the next eviction candidate: ```json { @@ -411,21 +411,29 @@ Example payload with one running node and one evicted node: "version": "0.4.0", "status": "running", "pid": 48213, - "uptime_secs": 3600 + "uptime_secs": 7200 }, { "node_id": 2, "name": "antnode2", "version": "0.4.0", + "status": "running", + "pid": 48219, + "uptime_secs": 6600 + }, + { + "node_id": 3, + "name": "antnode3", + "version": "0.4.0", "status": "evicted", "eviction": { - "reason": "Low disk: 480 MiB free, evicted to reclaim space", + "reason": "Automatically evicted to reclaim disk space: only 480 MiB free on its partition. Its data directory was deleted, recovering ~2.00 GiB.", "evicted_at": 1720800000, "reclaimed_bytes": 2147483648 } } ], - "total_running": 1, + "total_running": 2, "total_stopped": 1, "health": { "overall": "warning", @@ -433,13 +441,13 @@ Example payload with one running node and one evicted node: { "kind": "disk_space", "level": "warning", - "summary": "900 MiB free on /; node 1 is the next eviction candidate", - "partition": "/", + "summary": "Disk space low on dev:2049: 900 MiB free. An eviction may occur once it reaches 500 MiB; node 2 would be evicted next.", + "partition": "dev:2049", "available_bytes": 943718400, "eviction_threshold_bytes": 524288000, "candidate": { - "node_id": 1, - "data_dir": "/home/alice/.local/share/autonomi/node/antnode1", + "node_id": 2, + "data_dir": "/home/alice/.local/share/autonomi/node/antnode2", "size_bytes": 1073741824 } } @@ -452,14 +460,14 @@ Example payload with one running node and one evicted node: **Command:** `ant node dismiss ` -Removes a node's registry entry so it no longer appears in `ant node status`. This is the recovery step for a node the daemon evicted for low disk, though it is not restricted to evicted nodes. +Removes a node's registry entry so it no longer appears in `ant node status`. This clears the record for a node the daemon evicted for low disk, though it is not restricted to evicted nodes. -Dismissal behaviour depends on the daemon: +Dismissal behavior depends on the daemon: -- With the daemon running, dismiss removes any node that is not currently running. It refuses to dismiss a running node and asks you to stop it first. +- With the daemon running, dismiss removes any node that is not running. It refuses to dismiss a running node and asks you to stop it first. - With the daemon stopped, dismiss removes the registry entry directly and does not stop any process. Dismiss a node only when it is not running, so you do not leave an orphaned node process behind. -To recover capacity after an eviction, dismiss the retained `Evicted` record, then add a replacement node with `ant node add` if you still need it. +Dismissing removes the registry entry only; it does not itself reclaim disk space. After an eviction, check the eviction reason and `reclaimed_bytes` first: if the data directory was not deleted, free that space manually. Once enough capacity is available, dismiss the `Evicted` record, then add a replacement node with `ant node add` if you still need it. **Parameters:** @@ -473,6 +481,12 @@ To recover capacity after an eviction, dismiss the retained `Evicted` record, th ant node dismiss 3 ``` +Output: + +```text +✓ Dismissed node 3 (antnode3) +``` + ### `ant node stop` Stops all registered nodes, or one named node with `--service-name`. From b6751f0ce2855054d79908e2745c6fe0ee1a5be0 Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 13 Jul 2026 12:31:41 +0000 Subject: [PATCH 4/5] prose: use real node service names/paths, clarify reclaimed_bytes, name the node daemon Address review of the CLI command reference: - Use the CLI-generated identifiers in the status JSON example and dismiss output: service names node1/node2/node3 and the default data directory .../nodes/node-2 (ant-core registry assigns node{id}; node_data_dir builds /nodes/node-{id}). Dismiss output is now the audited "Dismissed node 3 (node3)". - Clarify that reclaimed_bytes is the recorded pre-deletion size estimate for the cleanup attempt, not a measurement of free space now available; direct readers to check actual free disk space before adding a replacement node. - Replace the ambiguous bare "the daemon" with "node daemon" in the new prose to name the interface explicitly, matching the node daemon sections. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_018k8QXTGkcfKXvycW1CAwRn --- docs/cli/command-reference.md | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/docs/cli/command-reference.md b/docs/cli/command-reference.md index 9272058..bf455a6 100644 --- a/docs/cli/command-reference.md +++ b/docs/cli/command-reference.md @@ -337,9 +337,9 @@ ant node start --service-name node1 ### `ant node status` -Shows the status of all registered nodes. Each node reports a state, shown in the table as `Running`, `Stopped`, `Starting`, `Stopping`, `Errored`, or `Evicted`. An evicted node is one the daemon automatically stopped when its host ran low on disk. Eviction stops the node and attempts to delete its data directory to reclaim space, then keeps the node's registry record marked `Evicted`. The table row for an evicted node shows the eviction reason and the exact `ant node dismiss` command to clear it. Deletion can fail; check the eviction reason, and `reclaimed_bytes` in the JSON output, to confirm whether space was actually recovered. +Shows the status of all registered nodes. Each node reports a state, shown in the table as `Running`, `Stopped`, `Starting`, `Stopping`, `Errored`, or `Evicted`. An evicted node is one the node daemon automatically stopped when its host ran low on disk. Eviction stops the node and attempts to delete its data directory to reclaim space, then keeps the node's registry record marked `Evicted`. The table row for an evicted node shows the eviction reason and the exact `ant node dismiss` command to clear it. Deletion can fail; check the eviction reason to see what was attempted, and treat `reclaimed_bytes` in the JSON output as the recorded estimate for that attempt rather than a measurement of free space now available. -When the daemon is running and a health snapshot is available, the output opens with a fleet-health summary of `Healthy`, `Warning`, or `Critical`, followed by one line per check that is not healthy. The summary is omitted when the daemon is stopped, and also when the daemon is running but the snapshot cannot be retrieved. +When the node daemon is running and a health snapshot is available, the output opens with a fleet-health summary of `Healthy`, `Warning`, or `Critical`, followed by one line per check that is not healthy. The summary is omitted when the node daemon is stopped, and also when it is running but the snapshot cannot be retrieved. **Parameters:** @@ -366,7 +366,7 @@ The payload has these top-level fields: | `nodes` | array | One object per registered node. | | `total_running` | integer | Count of nodes reported as `running` or `starting`. | | `total_stopped` | integer | Count of every other node, including `stopped`, `stopping`, `errored`, and `evicted`. | -| `health` | object or null | Fleet-health snapshot. It is `null` when the daemon is stopped, and also when the daemon is running but the snapshot cannot be retrieved, so `null` alone does not mean the daemon is stopped. | +| `health` | object or null | Fleet-health snapshot. It is `null` when the node daemon is stopped, and also when it is running but the snapshot cannot be retrieved, so `null` alone does not mean the node daemon is stopped. | Each `nodes` entry has these fields: @@ -407,7 +407,7 @@ Example payload with two running nodes that share a partition and one previously "nodes": [ { "node_id": 1, - "name": "antnode1", + "name": "node1", "version": "0.4.0", "status": "running", "pid": 48213, @@ -415,7 +415,7 @@ Example payload with two running nodes that share a partition and one previously }, { "node_id": 2, - "name": "antnode2", + "name": "node2", "version": "0.4.0", "status": "running", "pid": 48219, @@ -423,7 +423,7 @@ Example payload with two running nodes that share a partition and one previously }, { "node_id": 3, - "name": "antnode3", + "name": "node3", "version": "0.4.0", "status": "evicted", "eviction": { @@ -447,7 +447,7 @@ Example payload with two running nodes that share a partition and one previously "eviction_threshold_bytes": 524288000, "candidate": { "node_id": 2, - "data_dir": "/home/alice/.local/share/autonomi/node/antnode2", + "data_dir": "/home/alice/.local/share/autonomi/nodes/node-2", "size_bytes": 1073741824 } } @@ -460,14 +460,14 @@ Example payload with two running nodes that share a partition and one previously **Command:** `ant node dismiss ` -Removes a node's registry entry so it no longer appears in `ant node status`. This clears the record for a node the daemon evicted for low disk, though it is not restricted to evicted nodes. +Removes a node's registry entry so it no longer appears in `ant node status`. This clears the record for a node the node daemon evicted for low disk, though it is not restricted to evicted nodes. -Dismissal behavior depends on the daemon: +Dismissal behavior depends on the node daemon: -- With the daemon running, dismiss removes any node that is not running. It refuses to dismiss a running node and asks you to stop it first. -- With the daemon stopped, dismiss removes the registry entry directly and does not stop any process. Dismiss a node only when it is not running, so you do not leave an orphaned node process behind. +- When the node daemon is running, dismiss removes any node that is not running. It refuses to dismiss a running node and asks you to stop it first. +- When the node daemon is stopped, dismiss removes the registry entry directly and does not stop any process. Dismiss a node only when it is not running, so you do not leave an orphaned node process behind. -Dismissing removes the registry entry only; it does not itself reclaim disk space. After an eviction, check the eviction reason and `reclaimed_bytes` first: if the data directory was not deleted, free that space manually. Once enough capacity is available, dismiss the `Evicted` record, then add a replacement node with `ant node add` if you still need it. +Dismissing removes the registry entry only; it does not itself reclaim disk space. After an eviction, read the eviction reason to see whether cleanup succeeded, and treat `reclaimed_bytes` as the recorded estimate for that attempt, not proof of current free space. Before adding a replacement node, check the actual free disk space on the affected partition and free space manually if the eviction did not. Once enough capacity is available, dismiss the `Evicted` record, then add a replacement node with `ant node add` if you still need it. **Parameters:** @@ -484,7 +484,7 @@ ant node dismiss 3 Output: ```text -✓ Dismissed node 3 (antnode3) +✓ Dismissed node 3 (node3) ``` ### `ant node stop` From 6f67b08733bce7fdf134543fa6b011745e75fcab Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 13 Jul 2026 13:00:05 +0000 Subject: [PATCH 5/5] prose: correct default data_dir root to ant in node status JSON example config::data_dir() resolves to ~/.local/share/ant on Linux (config.rs), so the CLI-generated default node data directory is /ant/nodes/node-{id}. Fix the example candidate path from .../autonomi/nodes/node-2 to .../ant/nodes/node-2. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_018k8QXTGkcfKXvycW1CAwRn --- docs/cli/command-reference.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/cli/command-reference.md b/docs/cli/command-reference.md index bf455a6..38e34a8 100644 --- a/docs/cli/command-reference.md +++ b/docs/cli/command-reference.md @@ -447,7 +447,7 @@ Example payload with two running nodes that share a partition and one previously "eviction_threshold_bytes": 524288000, "candidate": { "node_id": 2, - "data_dir": "/home/alice/.local/share/autonomi/nodes/node-2", + "data_dir": "/home/alice/.local/share/ant/nodes/node-2", "size_bytes": 1073741824 } }