This message was deleted.
# harvester
a
This message was deleted.
r
run docker pull from another vm (on the same harvester cluster) didn’t have issue. we only observed issue when pull images from outside.
g
that is hard to say.. likely something with harbor? anything useful in the harbor logs?
r
maybe not related to harbor and harvester, we observed some large file transmission was truncated, so the docker client is keep waiting.
g
ok
r
and this situation only happened if the request are not from the same network. like request from aws instance can experience this situation.
we are suspecting maybe the underlying network configuration issue
g
ok
r
I think this would be an issue, my harvester default mtu is 1500 but the underlying configurtation only use 1450
g
You can customise mtu when setting up the node
r
are we able to mdify it on-the-fly?
g
I need to check