# Can not realize distributed training across machines with DDP

**URL:** <https://discuss.pytorch.org/t/can-not-realize-distributed-training-across-machines-with-ddp/144519>\
**Category:** distributed\
**Created:** [February 19, 2022, 7:02am UTC](https://discuss.pytorch.org/t/can-not-realize-distributed-training-across-machines-with-ddp/144519 "2022-02-19T07:02:34Z")\
**Posts on this page:** 1\
**Showing post:** 8

<div class="post-metadata">

**Author:** ![taehyunzzz](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/taehyunzzz/32/42473_2.png) [@taehyunzzz](https://discuss.pytorch.org/u/taehyunzzz)\
**Post date:** [March 2, 2022, 5:10pm UTC](https://discuss.pytorch.org/t/can-not-realize-distributed-training-across-machines-with-ddp/144519/8 "2022-03-02T17:10:54Z")

</div>

I think the problem is your firewall. Try [this](https://discuss.pytorch.org/t/connection-refused-with-gloo-process-group-initialization/144857/6).

---

_[View the full topic](https://discuss.pytorch.org/t/can-not-realize-distributed-training-across-machines-with-ddp/144519)._