Skip to content

GitLab

Unverified Commit cfd302db authored by Or Ozeri's avatar Or Ozeri Committed by GitHub
Browse files

OffloadingConnector: Fix GPU block tracking bug (#25856) 


Signed-off-by: default avatarOr Ozeri <oro@il.ibm.com>
parent fb610ae6
Hide whitespace changes
Inline Side-by-side
Showing with 3 additions and 2 deletions
vllm/distributed/kv_transfer/kv_connector/v1/offloading_connector.py
View file @ cfd302db
...@@ -278,8 +278,9 @@ class OffloadingConnectorScheduler: ...@@ -278,8 +278,9 @@ class OffloadingConnectorScheduler:
req, start_idx=start_block_idx, end_idx=num_blocks) req, start_idx=start_block_idx, end_idx=num_blocks)
store_output = self.manager.prepare_store(new_block_hashes) store_output = self.manager.prepare_store(new_block_hashes)
if store_output is None: if store_output is None:
logger.warning("Cannot store %s blocks", num_new_blocks) logger.warning("Request %s: cannot store %s blocks", req_id,
break num_new_blocks)
continue
self._next_stored_block_idx[req_id] = num_blocks self._next_stored_block_idx[req_id] = num_blocks
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment