Jiaxin Shan
Jiaxin Shan
@XiaobinZhao can you try to use `--dist-init-addr "${PODSET_NAME}-0.${STORM_SERVICE_NAME}.default.svc.cluster.local:5000" \` for service discovery. the latest router has filter pod with `stormservice.orchestration.aibrix.ai/pod-group-index=0` pods. Could I know the router and controller manager version...
@XiaobinZhao you need to use nightly aibrix image. replace v0.4.1 with `nightly` image. v0.4.1 doesn't have all the features BTW, I update the above yaml. (use $POD_GROUP_INDEX to replace $ROLE_REPLICA_INDEX)....
@varungup90 @zhangjyr 1. Technically, this is good for debugging but has too many limitations. I suggest we hold this one at this moment and see other ways. I think ask...
vllm need ray cluster to cross-node, we need to launch ray inside the vllm to use PodGroupSize. are you able to use sglang at this moment?
/cc @DwyaneShi
@xvoron thanks for the feedback. We will take a look at this problem and address #1014 together. /cc @zhangjyr
Sure. I think we just need to add a vLLM multi-node example using stormservice. is that what you are looking for?
@ying2025 got you. I will ask some contributors to write some concrete examples and explanation. the podGroupSize is a little bit confusing to most users. We will make sure this...
We recommend users to adopt helm way to install aibrix now. CRDs separation could be separated and this is out-of-box supported. The tricky thing is still to support dependencies. @Xunzhuo...
@yang753 yes. v0.4.1 replicas is not supported. BTW, I happen to refactor this part. i want to double check with the behaviors could you check this PR https://github.com/vllm-project/aibrix/pull/1670? the new...