Compare commits

...

4 Commits

Author SHA1 Message Date
Alex Cheema
c84b565a32 feat: disable RDMA instance type when no Thunderbolt 5 hardware detected
When the cluster has fewer than 2 nodes with Thunderbolt 5 hardware,
the MLX RDMA button is now disabled with a tooltip explaining that
RDMA requires Thunderbolt 5. This prevents users from selecting an
instance type that will fail at placement time.

Closes #1351

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-17 10:01:24 -08:00
Alex Cheema
db79c350c1 Fix graceful process shutdown in macOS app (#1372)
## Motivation

Fixes #1370

When the macOS app stops exo, GPU/system memory isn't released. This
happens because:

1. The macOS app calls `process.terminate()` (SIGTERM) but the Python
process only registers a graceful shutdown handler for SIGINT, not
SIGTERM. SIGTERM's default Python behavior raises `SystemExit` which
bypasses the cleanup cascade (runner subprocess MLX cleanup via
`mx.clear_cache()`, channel closing, etc.).
2. The app doesn't wait for the process to actually finish cleanup — it
immediately nils out the process reference.

## Changes

**`src/exo/main.py`**: Register SIGTERM handler alongside SIGINT so the
graceful shutdown cascade (`Node.shutdown()` → cancel task group →
worker/runner cleanup → `mx.clear_cache()` + `gc.collect()`) runs
regardless of which signal is received.

**`app/EXO/EXO/ExoProcessController.swift`**: Replace immediate
`process.terminate()` with escalating shutdown per @Evanev7's
suggestion:
1. Send SIGINT via `process.interrupt()` — triggers the registered
Python handler for graceful cleanup
2. Wait up to 5 seconds for the process to exit
3. If still running, escalate to SIGTERM via `process.terminate()`
4. Wait up to 3 seconds
5. If still running, force kill via SIGKILL

The escalation runs in a detached `Task` so the UI updates immediately
(status → stopped) without blocking.

## Why It Works

The root cause is that SIGTERM wasn't triggering the graceful shutdown
path. By registering a SIGTERM handler in Python and sending SIGINT
first from the macOS app, the process gets a chance to run the full
cleanup cascade: cancelling the task group, shutting down runners (which
call `del model; mx.clear_cache(); gc.collect()`), closing channels, and
flushing logs. The escalation to SIGTERM and SIGKILL ensures the process
always terminates even if graceful shutdown hangs.

## Test Plan

### Manual Testing
<!-- Hardware: Mac Studio M4 Max 128GB -->
- Start exo via macOS app, load a model, run inference
- Stop via the toggle switch, verify memory is released without
requiring a system restart
- Test rapid stop/start (restart) to ensure no race conditions

### Automated Testing
- `uv run basedpyright` — 0 errors
- `uv run ruff check` — passes
- `nix fmt` — no changes

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: Evan Quiney <evanev7@gmail.com>
2026-02-17 09:03:54 -08:00
Alex Cheema
d6301ed593 dashboard: redesign downloads page as model×node table (#1465)
## Motivation

The current downloads page uses a node-centric card grid layout that is
messy and hard to read — the same model across different nodes appears
in separate cards, and deep nesting wastes space. This makes it
difficult to quickly see which models are on which nodes.

## Changes

Rewrote the downloads page
(`dashboard/src/routes/downloads/+page.svelte`) from a card grid to a
clean table layout:

- **Rows** = models (unique across all nodes)
- **Columns** = nodes (with disk free shown in header)
- **Cells** show status at a glance:
  -  Green checkmark + size for completed downloads
  - 🟡 Yellow percentage + mini progress bar + speed for active downloads
  - `...` for pending downloads
  -  Red X for failed downloads
  - `--` for models not present on a node
- Delete/download action buttons appear on row hover
- Model name column is sticky on horizontal scroll (for many-node
clusters)
- Models sorted by number of nodes with completed downloads
- Imported shared utilities from `$lib/utils/downloads` instead of
inline re-implementations

### Backend: model directory in download events

- Added `model_directory` field to `BaseDownloadProgress` so all
download status events include the on-disk path
- Added `_model_dir()` helper to `DownloadCoordinator` to compute the
path from `EXO_MODELS_DIR`
- Dashboard uses this to show file location and enable "open in Finder"
for completed downloads

### Info modal

- Clicking a model name opens an info modal showing card details
(family, quantization, capabilities, storage size, layer count, tensor
parallelism support)

### Other fixes

- Fixed model name truncation in the table
- Excluded `tests/start_distributed_test.py` from pytest collection (CLI
script that calls `sys.exit()` at import time)

## Test Plan

- [x] `uv run basedpyright` — 0 errors
- [x] `uv run ruff check` — all passed
- [x] `nix fmt` — clean
- [x] `uv run pytest` — 188 passed, 1 skipped

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-17 14:31:47 +00:00
Evan Quiney
6d1ca6689b don't time out node identities (#1493)
currently nodes leaving and rejoining the cluster can lose their identity. We have no need to delete this data on node timing out, so let's just persist it.
2026-02-17 11:48:28 +00:00
8 changed files with 640 additions and 500 deletions

View File

@@ -126,11 +126,37 @@ final class ExoProcessController: ObservableObject {
return
}
process.terminationHandler = nil
if process.isRunning {
process.terminate()
}
self.process = nil
status = .stopped
guard process.isRunning else {
self.process = nil
return
}
let proc = process
self.process = nil
Task.detached {
proc.interrupt()
for _ in 0..<50 {
if !proc.isRunning { return }
try? await Task.sleep(nanoseconds: 100_000_000)
}
if proc.isRunning {
proc.terminate()
}
for _ in 0..<30 {
if !proc.isRunning { return }
try? await Task.sleep(nanoseconds: 100_000_000)
}
if proc.isRunning {
kill(proc.processIdentifier, SIGKILL)
}
}
}
func restart() {

View File

@@ -114,6 +114,16 @@
});
let tb5InfoDismissed = $state(false);
// Whether the cluster has >=2 nodes with RDMA-capable (TB5) hardware
const hasRdmaCapableNodes = $derived.by(() => {
const ids = tbIdentifiers;
if (!ids) return false;
const tb5NodeCount = Object.values(ids).filter(
(node) => node.interfaces.length > 0,
).length;
return tb5NodeCount >= 2;
});
// Helper to get friendly node name from node ID
function getNodeName(nodeId: string): string {
const node = data?.nodes?.[nodeId];
@@ -2882,28 +2892,48 @@
</span>
MLX Ring
</button>
<button
onclick={() => {
selectedInstanceType = "MlxIbv";
saveLaunchDefaults();
}}
class="flex items-center gap-2 py-2 px-4 text-sm font-mono border rounded transition-all duration-200 cursor-pointer {selectedInstanceType ===
'MlxIbv'
? 'bg-transparent text-exo-yellow border-exo-yellow'
: 'bg-transparent text-white/70 border-exo-medium-gray/50 hover:border-exo-yellow/50'}"
>
<span
class="w-4 h-4 rounded-full border-2 flex items-center justify-center {selectedInstanceType ===
'MlxIbv'
? 'border-exo-yellow'
: 'border-exo-medium-gray'}"
<div class="group relative">
<button
onclick={() => {
if (!hasRdmaCapableNodes) return;
selectedInstanceType = "MlxIbv";
saveLaunchDefaults();
}}
disabled={!hasRdmaCapableNodes}
class="flex items-center gap-2 py-2 px-4 text-sm font-mono border rounded transition-all duration-200 {hasRdmaCapableNodes
? 'cursor-pointer'
: 'cursor-not-allowed opacity-40'} {selectedInstanceType ===
'MlxIbv' && hasRdmaCapableNodes
? 'bg-transparent text-exo-yellow border-exo-yellow'
: 'bg-transparent text-white/70 border-exo-medium-gray/50'} {hasRdmaCapableNodes &&
selectedInstanceType !== 'MlxIbv'
? 'hover:border-exo-yellow/50'
: ''}"
>
{#if selectedInstanceType === "MlxIbv"}
<span class="w-2 h-2 rounded-full bg-exo-yellow"></span>
{/if}
</span>
MLX RDMA
</button>
<span
class="w-4 h-4 rounded-full border-2 flex items-center justify-center {selectedInstanceType ===
'MlxIbv' && hasRdmaCapableNodes
? 'border-exo-yellow'
: 'border-exo-medium-gray'}"
>
{#if selectedInstanceType === "MlxIbv" && hasRdmaCapableNodes}
<span class="w-2 h-2 rounded-full bg-exo-yellow"
></span>
{/if}
</span>
MLX RDMA
</button>
{#if !hasRdmaCapableNodes}
<div
class="absolute top-full left-0 mt-2 w-64 p-3 rounded border border-blue-400/30 bg-exo-dark-gray/95 backdrop-blur-sm opacity-0 invisible group-hover:opacity-100 group-hover:visible transition-all duration-200 z-50 shadow-lg"
>
<p class="text-xs text-white/80">
RDMA requires Thunderbolt 5 hardware on at least 2
nodes. Your devices do not have Thunderbolt 5.
</p>
</div>
{/if}
</div>
</div>
</div>

View File

File diff suppressed because it is too large Load Diff

View File

@@ -132,7 +132,7 @@ markers = [
env = [
"EXO_TESTS=1"
]
addopts = "-m 'not slow'"
addopts = "-m 'not slow' --ignore=tests/start_distributed_test.py"
filterwarnings = [
"ignore:builtin type Swig:DeprecationWarning",
]

View File

@@ -14,6 +14,7 @@ from exo.download.download_utils import (
map_repo_download_progress_to_download_progress_data,
)
from exo.download.shard_downloader import ShardDownloader
from exo.shared.constants import EXO_MODELS_DIR
from exo.shared.models.model_cards import ModelId
from exo.shared.types.commands import (
CancelDownload,
@@ -63,6 +64,9 @@ class DownloadCoordinator:
self.event_sender, self.event_receiver = channel[Event]()
self.shard_downloader.on_progress(self._download_progress_callback)
def _model_dir(self, model_id: ModelId) -> str:
return str(EXO_MODELS_DIR / model_id.normalize())
async def _download_progress_callback(
self, callback_shard: ShardMetadata, progress: RepoDownloadProgress
) -> None:
@@ -74,6 +78,7 @@ class DownloadCoordinator:
shard_metadata=callback_shard,
node_id=self.node_id,
total_bytes=progress.total_bytes,
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = completed
await self.event_sender.send(
@@ -93,6 +98,7 @@ class DownloadCoordinator:
download_progress=map_repo_download_progress_to_download_progress_data(
progress
),
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = ongoing
await self.event_sender.send(
@@ -170,7 +176,11 @@ class DownloadCoordinator:
return
# Emit pending status
progress = DownloadPending(shard_metadata=shard, node_id=self.node_id)
progress = DownloadPending(
shard_metadata=shard,
node_id=self.node_id,
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = progress
await self.event_sender.send(NodeDownloadProgress(download_progress=progress))
@@ -184,6 +194,7 @@ class DownloadCoordinator:
shard_metadata=shard,
node_id=self.node_id,
total_bytes=initial_progress.total_bytes,
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = completed
await self.event_sender.send(
@@ -206,6 +217,7 @@ class DownloadCoordinator:
download_progress=map_repo_download_progress_to_download_progress_data(
initial_progress
),
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = status
self.event_sender.send_nowait(NodeDownloadProgress(download_progress=status))
@@ -219,6 +231,7 @@ class DownloadCoordinator:
shard_metadata=shard,
node_id=self.node_id,
error_message=str(e),
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = failed
await self.event_sender.send(
@@ -253,6 +266,7 @@ class DownloadCoordinator:
pending = DownloadPending(
shard_metadata=current_status.shard_metadata,
node_id=self.node_id,
model_directory=self._model_dir(model_id),
)
await self.event_sender.send(
NodeDownloadProgress(download_progress=pending)
@@ -295,11 +309,18 @@ class DownloadCoordinator:
node_id=self.node_id,
shard_metadata=progress.shard,
total_bytes=progress.total_bytes,
model_directory=self._model_dir(
progress.shard.model_card.model_id
),
)
elif progress.status in ["in_progress", "not_started"]:
if progress.downloaded_bytes_this_session.in_bytes == 0:
status = DownloadPending(
node_id=self.node_id, shard_metadata=progress.shard
node_id=self.node_id,
shard_metadata=progress.shard,
model_directory=self._model_dir(
progress.shard.model_card.model_id
),
)
else:
status = DownloadOngoing(
@@ -308,6 +329,9 @@ class DownloadCoordinator:
download_progress=map_repo_download_progress_to_download_progress_data(
progress
),
model_directory=self._model_dir(
progress.shard.model_card.model_id
),
)
else:
continue

View File

@@ -136,6 +136,8 @@ class Node:
async def run(self):
async with self._tg as tg:
signal.signal(signal.SIGINT, lambda _, __: self.shutdown())
signal.signal(signal.SIGTERM, lambda _, __: self.shutdown())
tg.start_soon(self.router.run)
tg.start_soon(self.election.run)
if self.download_coordinator:
@@ -147,8 +149,6 @@ class Node:
if self.api:
tg.start_soon(self.api.run)
tg.start_soon(self._elect_loop)
signal.signal(signal.SIGINT, lambda _, __: self.shutdown())
signal.signal(signal.SIGTERM, lambda _, __: self.shutdown())
def shutdown(self):
# if this is our second call to shutdown, just sys.exit

View File

@@ -218,11 +218,6 @@ def apply_node_timed_out(event: NodeTimedOut, state: State) -> State:
key: value for key, value in state.downloads.items() if key != event.node_id
}
# Clean up all granular node mappings
node_identities = {
key: value
for key, value in state.node_identities.items()
if key != event.node_id
}
node_memory = {
key: value for key, value in state.node_memory.items() if key != event.node_id
}
@@ -263,7 +258,6 @@ def apply_node_timed_out(event: NodeTimedOut, state: State) -> State:
"downloads": downloads,
"topology": topology,
"last_seen": last_seen,
"node_identities": node_identities,
"node_memory": node_memory,
"node_disk": node_disk,
"node_system": node_system,

View File

@@ -26,6 +26,7 @@ class DownloadProgressData(CamelCaseModel):
class BaseDownloadProgress(TaggedModel):
node_id: NodeId
shard_metadata: ShardMetadata
model_directory: str = ""
class DownloadPending(BaseDownloadProgress):