Closed LeoHChen closed 4 years ago
@LeoHChen just to notify.. there are a lot of users that still use node.sh and didnt/dont want to go to using the binary yet. Wht is the solution for them, other than going to the binary?
all validator node(except explorer and leader) already change to use binary on AWS and GCP. will change explorer node now
all explorer node already change to use binary
pushed change node to start by binary Ansible script. see https://github.com/harmony-one/ansible/pull/6
@LeoHChen just to notify.. there are a lot of users that still use node.sh and didnt/dont want to go to using the binary yet. Wht is the solution for them, other than going to the binary?
node.sh will be depreciated. @mindstyle85 thanks for the inquiry. Can you lead a community discussion and proposal on the retirement of node.sh?
looks good @jhd2best , have you made changes on the terraform scripts?
pushed change node to start by binary Terraform script. see https://github.com/harmony-one/terraform/pull/5
nice work
seems everything is stable. we can do the leader update tonight. @jhd2best
Fixed node.sh
, https://github.com/harmony-one/harmony/pull/3403
node.sh deprecation proposal https://talk.harmony.one/t/node-sh-deprecation/949
all leader nodes changed successful
myip.opendns.com is not available to use now, it will stop node.sh from running due to the ip checking failure.
Now it is time to stop using node.sh on harmony's mainnet nodes. Instead, we can just run
harmony
binary directory in systemd service.The command to use is
harmony --run.legacy --http.ip 0.0.0.0
Please update all harmony.service file on all nodes, to use harmony binary directly.
Also, our terraform scripts need to be updated for this change as well.
Noted, on GCP node, I've noticed sometimes harmony is failed to start in systemd because the selinux prevent the user process from running. Need to disable selinux permanently on those nodes and disable selinux in current running process.