Our team has resolved the Goerli head stall for some node operators, and public infrastructure should now be fully operational for all users.

As a precaution, we advise all OP Stack node operators (including mainnet chains) to change their op-node to use ws instead of http for the l2 connection to op-geth while we're investigating the underlying issue further. So if you use --l2 http://localhost:8551 to connect to op-geth, change that to --l2 ws://localhost:8551. No need to restart op-geth. We don't expect the issue that affected some Goerli nodes to be a risk to any mainnet OP Stack nodes, but still advise node operators to change this setting as a precaution.

If you continue to experience an issue, please let us know on Discord.

Thanks for your patience with us!

Our team has identified the cause of the Goerli head stall for some node operators, and is actively working on a fix.

At this time, some users of public infrastructure and node operators will continue to experience a head stall.

We advise node operators to change their op-node to use ws instead of http for the l2 connection to op-geth as a quick remedy while we're working on a proper fix. So if you use --l2 http://localhost:8551 to connect to op-geth, change that to --l2 ws://localhost:8551. No need to restart op-geth.

We’ll continue to share updates as we have them.