You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* feat: enable multi-datacenter support
* docs: update SDK reference and deploy guide for multi-datacenter support
* feat: add datacenter alias on NetworkVolume
* fix: pass locations through manifest and resource provisioner
* fix: remove dead _is_prod_environment function
* fix: address copilot review on multi-volume support
* fix: address PR review comments for multi-datacenter support
* style: fix ruff formatting in manifest.py
* fix: remove redundant type annotation on _deployed_volume_ids reset
* fix: align datacenter enum to s3-enabled DCs
* fix: address post-rebase review comments
* fix: address PR review findings for multi-datacenter support
- fix US_GA_1 references to US_GA_2 in docs, tests (enum member does not exist)
- add field_serializer for datacenter on ServerlessResource for consistency
- strip runtime-assigned volume IDs from config_hash to prevent false drift
- add tests for multi-volume manifest serialization
- add tests for multi-volume provisioner reconstruction
- add test verifying volume id does not affect config hash
|`volume`|`NetworkVolume`or list |`None`| Network volume(s) for persistent storage. One volume per datacenter. |
50
+
|`datacenter`|`DataCenter`, list, `str`, or `None`|`None`|Datacenter(s) to deploy into. `None` means all available DCs. Accepts a single value, a list, or string DC IDs. CPU endpoints must use DCs in `CPU_DATACENTERS`. |
51
51
|`env`|`dict[str, str]`|`None`| Environment variables for the endpoint. |
52
52
|`gpu_count`|`int`|`1`| GPUs per worker. |
53
53
|`execution_timeout_ms`|`int`|`0`| Max execution time in ms. 0 = no limit. |
@@ -335,8 +335,25 @@ CPU instance selection. Can also be passed as a string to `cpu=`.
335
335
336
336
| Value | Location |
337
337
|-------|----------|
338
-
|`DataCenter.EU_RO_1`| Europe - Romania (default) |
339
-
|`DataCenter.US_TX_3`| US - Texas |
338
+
|`DataCenter.US_CA_2`| US - California |
339
+
|`DataCenter.US_GA_2`| US - Georgia |
340
+
|`DataCenter.US_IL_1`| US - Illinois |
341
+
|`DataCenter.US_KS_2`| US - Kansas |
342
+
|`DataCenter.US_MD_1`| US - Maryland |
343
+
|`DataCenter.US_MO_1`| US - Missouri |
344
+
|`DataCenter.US_MO_2`| US - Missouri |
345
+
|`DataCenter.US_NC_1`| US - North Carolina |
346
+
|`DataCenter.US_NC_2`| US - North Carolina |
347
+
|`DataCenter.US_NE_1`| US - Nebraska |
348
+
|`DataCenter.US_WA_1`| US - Washington |
349
+
|`DataCenter.EU_CZ_1`| Europe - Czech Republic |
350
+
|`DataCenter.EU_RO_1`| Europe - Romania |
351
+
|`DataCenter.EUR_IS_1`| Europe - Iceland |
352
+
|`DataCenter.EUR_NO_1`| Europe - Norway |
353
+
354
+
When `datacenter=None` (the default), the endpoint is available in all data centers.
355
+
356
+
CPU endpoints are restricted to the `CPU_DATACENTERS` subset: `EU_RO_1`.
340
357
341
358
### CudaVersion
342
359
@@ -350,12 +367,22 @@ CPU instance selection. Can also be passed as a string to `cpu=`.
350
367
351
368
### NetworkVolume
352
369
353
-
Persistent storage that survives worker restarts.
370
+
Persistent storage that survives worker restarts. Each volume is tied to a specific datacenter.
354
371
355
372
```python
356
-
from runpod_flash import NetworkVolume
373
+
from runpod_flash import NetworkVolume, DataCenter
0 commit comments