Data from PoolTool showing around 45% of nodes still need updating to 1.20.0
Cardano-Node 1.20.0 update involves a database migration that takes around 10-30 mins to complete. To reduce downtime, it is possible to create a snapshot of the migrated database and apply it to subsequent nodes needing updates.
Agree, thanks for the videos. You really are helping the community!
Can you explain the sequence/order of what (BP, relay, or relay then BP) should be updated first to minimize downtime. More importantly, if done right the objective is to do the update without losing sync/connectivity with the blockchain correct? In other words, if you are a slot leader and do this update you will not produce a block, correct, if the BP node is getting updated during the slot…? Or can you keep the blockchain happy by temporarily processing transactions with only a relay??
So is there a way to keep the blocks coming and have no downtime? Or is downtime required?
In the future, when a stake pool is saturated there will be a high probability of them being a slot leader for several ( a lot!!) of block so when are these updates suppose to occur if a down node being down for maintenance will not produce blocks?
Thanks in advance.
And as always, cheers and beers!
Your understanding is correct - the BP must always be up and during the update, it will be down and unable to produce blocks. In order to have no downtime, while the BP is updating, you have to run your BP on another node - usually a backup BP or if you want, you can transfer your pool keys to a relay and turn that relay into a BP while your main BP is updating.
Hope that’s clear enough Let me know if you want more details
Thanks for the clear explanation. That sounds easy but Im sure there are a lot of pitfalls and risks involved in transferring BP keys temporarily to a relay. When I was building the BP and relays I recall there were more scripts and commands involved for the BP. Were these extra scripts/commands only for the private keys and everything else “under the hood” for the engine is same between BP and relay?? In other words is there only a transfer of keys that occurs? Do these keys include the pool cert, delegation cert (all keys and certs) that tell the blockchain that its the same BP (formerly node)?
You are right again, the software is exactly the same - the only difference is that the BP runs with the pool parameters (vrf and kes keys + operational certificate). The additional thing to consider if transferring a BP to what was previously a relay is topology - you will want to change that to point only to your relays rather than public relays as it was previously set up.