[P/D][PCP]bugfix pcp force free twice caused logger error (#6124)
### What this PR does / why we need it?
The issue of the D node mistakenly sending the pull-end signal twice,
leading to the P node printing logger errors abnormally, has been
resolved.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By ci
- vLLM version: v0.13.0
- vLLM main:
d68209402d
Signed-off-by: wangxiaoteng <wangxiaoteng@huawei.com>
This commit is contained in:
@@ -449,6 +449,8 @@ class KVCacheRecvingThread(threading.Thread):
|
|||||||
f"{remote_request_id}: {e}",
|
f"{remote_request_id}: {e}",
|
||||||
exc_info=True)
|
exc_info=True)
|
||||||
finally:
|
finally:
|
||||||
|
self._send_done_signal_to_free_remote_port(remote_request_id, remote_host,
|
||||||
|
remote_port_send_num)
|
||||||
if all_task_done:
|
if all_task_done:
|
||||||
self.task_tracker.update_done_task_count(request_id)
|
self.task_tracker.update_done_task_count(request_id)
|
||||||
if request_id in self.proc_not_transfer_request:
|
if request_id in self.proc_not_transfer_request:
|
||||||
@@ -460,8 +462,6 @@ class KVCacheRecvingThread(threading.Thread):
|
|||||||
self._send_done_recv_signal(remote_request_id, remote_host,
|
self._send_done_recv_signal(remote_request_id, remote_host,
|
||||||
remote_handshake_port,
|
remote_handshake_port,
|
||||||
remote_port_send_num)
|
remote_port_send_num)
|
||||||
self._send_done_signal_to_free_remote_port(remote_request_id, remote_host,
|
|
||||||
remote_port_send_num)
|
|
||||||
|
|
||||||
def _send_done_signal_to_free_remote_port(
|
def _send_done_signal_to_free_remote_port(
|
||||||
self, request_id: str, remote_host: str,
|
self, request_id: str, remote_host: str,
|
||||||
|
|||||||
Reference in New Issue
Block a user