Files
xc-llm-ascend/vllm_ascend
lty 082aa2e5b7 [Bugfix]The service fails to be started when the memcache pool is enabled (#6229)
### What this PR does / why we need it?
The service fails to be started when the memcache pool is enabled
without configuring the mooncake path.

### Does this PR introduce _any_ user-facing change?
NA

### How was this patch tested?
```
#memcache
echo 200000 > /proc/sys/vm/nr_hugepages
source /usr/local/memfabric_hybrid/set_env.sh
source /usr/local/memcache_hybrid/set_env.sh
source /usr/local/Ascend/ascend-toolkit/set_env.sh
source /usr/local/Ascend/nnal/atb/set_env.sh
export MMC_LOCAL_CONFIG_PATH=/usr/local/memcache_hybrid/latest/config/mmc-local.conf

vllm serve /mnt/weight/DeepSeek-V3.2-Exp-W8A8 \
  --host $local_ip \
  --port 8002 \
  --served-model-name model \
  --data-parallel-size 2 \
  --tensor-parallel-size 8 \
  --enable-expert-parallel \
  --no-enable-prefix-caching \
  --no-enable-chunked-prefill \
  --max-num-seqs 4 \
  --max-model-len 8192 \
  --max-num-batched-tokens 8192 \
  --gpu-memory-utilization 0.9 \
  --trust-remote-code \
  --enforce-eager \
  --quantization ascend \
  --additional_config '{"ascend_scheduler_config":{"enabled":false}}' \
  --kv-transfer-config \
    '{
            "kv_connector": "AscendStoreConnector",
            "kv_role": "kv_both",
            "kv_connector_extra_config": {
	            "backend": "memcache",
                "lookup_rpc_port":"0"
            }
    }'
```

- vLLM version: v0.14.0
- vLLM main:
d68209402d

---------

Signed-off-by: lty <linhebiwen@gmail.com>
2026-02-02 16:26:18 +08:00
..
2026-01-22 15:46:59 +08:00