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
bitcoind becoming unreachable potentially restarting and becoming reachable again
In the former case we can probably just poll for readiness on startup. It's unlikely that something urgent has to be done, so waiting for bitcoind to finish its startup before continuing ours seems reasonable (we couldn't do anything anyway). The second case is dangerous, we can't just wait for it to come back, it might not come back at all, and we might be in the middle of an urgent situation. We currently just die whenever we lose contact to bitcoind, but I'd like to ween off of the full node dependency slowly, which is why I'm implementing the bitcoin protocol and am planning to keep a connection pool to public nodes. That'd allow us to monitor for blocks and transactions, and it allows us to broadcast transactions even without a bitcoind running locally.
So if this is an attempt to fix the startup problem, I'm all for it. If it is more about fixing outages while running, I'd say we allow some downtime, but if bitcoind doesn't come back in time we should be looking for alternatives to settle and shut down.
Seems to go out to lunch on reorgs:
+136792.168286138 lightningd(9465):BROKEN: bitcoin-cli getchaintips exited 28: 'error code: -28
error message:
Rewinding blocks...
Closes: ElementsProject#286
Signed-off-by: Rusty Russell <[email protected]>
Seems to go out to lunch on reorgs:
+136792.168286138 lightningd(9465):BROKEN: bitcoin-cli getchaintips exited 28: 'error code: -28
error message:
Rewinding blocks...
Closes: ElementsProject#286
Signed-off-by: Rusty Russell <[email protected]>
From #284 crash.log:
This is unfortunate, but we should give it a few tries I think?
The text was updated successfully, but these errors were encountered: