You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
right now, proposer can continue producing blocks while da is down, except if the node is restarted, that will require a connection to a da light client synced to start producing blocks.
this maybe an issue in case there are disruptions in celestia network, because nodes will be able to continue producing blocks until max skew time, but only in case node is not restarted.
i think this flow could be improved, allowing nodes to restart even when da is not available. although we still need some safety checks on rollapp startup to make sure da is available for new rollapps to enable ibc bridging, which could be done from roller.
The text was updated successfully, but these errors were encountered:
Are you saying it's good to produce while DA is down? But I suppose, we don't want that to continue indefinitely? And you're also saying nodes should be able to start and start producing, even if DA is down?
we have the skew time to be used to limit the gap between produced and submitted.
right now its possible to continue producing blocks while da is not available, except on startup (sequencer waits for da to be ready).
so a sequencer can continue producing blocks when da is down (with the limit of max skew), but if its restarted it wont be able to do it. what i propose is to change the code that when the node is started it does not wait for da readiness and same skew applies, this way if a node is restarted but da is down it is possible to continue producing blocks.
right now, proposer can continue producing blocks while da is down, except if the node is restarted, that will require a connection to a da light client synced to start producing blocks.
this maybe an issue in case there are disruptions in celestia network, because nodes will be able to continue producing blocks until max skew time, but only in case node is not restarted.
i think this flow could be improved, allowing nodes to restart even when da is not available. although we still need some safety checks on rollapp startup to make sure da is available for new rollapps to enable ibc bridging, which could be done from roller.
The text was updated successfully, but these errors were encountered: