CalvinBackup/OBLITERATUS
1
0
Fork 0
mirror of https://github.com/elder-plinius/OBLITERATUS.git synced 2026-08-29 22:20:36 +02:00
Code Issues Packages Projects Releases Wiki Activity

feat: publish GPU model lifecycle events

Browse Source
This commit is contained in:
Joseph Magly
2026-08-24 20:27:21 -04:00
parent 7aae98a367
commit 4b677be612
5 changed files with 331 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/deployment/shared-gpu-host.md
+32 -3
View File
@@ -50,9 +50,38 @@ the requested devices and VRAM before OBLITERATUS initializes CUDA, maintain the
lease for the entire process, and release it after CUDA allocations terminate.
Schedulers that drain another inference service commonly distinguish `acquire`
from `ready`. OBLITERATUS does not yet emit a portable post-model-load readiness
signal. Such deployments must not mark a lease ready immediately after process
start; add and test an application readiness hook first.
from `ready`. Set `OBLITERATUS_GPU_LIFECYCLE_DIR` to an existing runtime
directory writable by the OBLITERATUS service account to enable the local
lifecycle protocol. When it is unset, publication is a no-op and conservative
wrappers should retain their lease for the full process lifetime.
The application atomically replaces `current.json` and appends ordered JSON
objects to `events.jsonl`. Schema version 1 publishes `loading`, `resize`,
`ready`, `heartbeat`, and `release`, with run/model identity, a monotonic
sequence, process ID, timestamp, and measured allocator bytes. Readiness occurs
after model allocation, never at UI startup. Set
`OBLITERATUS_GPU_HEARTBEAT_SECONDS` to change the default 15-second heartbeat.
A minimal systemd setup lets the supervisor create the local boundary without
giving the application scheduling authority:
```ini
[Service]
User=obliteratus
Group=obliteratus-gpu
RuntimeDirectory=obliteratus-gpu-lifecycle
RuntimeDirectoryMode=0770
Environment=OBLITERATUS_GPU_LIFECYCLE_DIR=/run/obliteratus-gpu-lifecycle
Environment=OBLITERATUS_GPU_HEARTBEAT_SECONDS=15
ExecStart=/srv/obliteratus/current/.venv/bin/obliteratus ui --host 127.0.0.1
```
The host supervisor should acquire capacity before starting this unit, consume
the runtime files as untrusted structured data, deduplicate by `event_id`, and
release or resize its reservation only after the corresponding application
event. It should treat a stale heartbeat or process exit as a failed lease and
must not pass its control socket, administrative API, or scheduling permissions
to the application account.
The requested VRAM must include weights, activation collection, verification,
checkpoint snapshots, CUDA context, and a safety margin. Multi-GPU sharding is