How often do you have to rebuild a node?

Staking & Delegation Operate a Stake Pool
1

First post here:
I’m currently thinking about implementing a stake pool, and try to have a clearer picture of what is to be done as maintenance, (beside server security issues).

I’m wondering how often do you have to rebuild a cardano-node. And what are the signals/events that make you decide a rebuild of a cardano-node ?

Moreover, (I have not figured out the difference between a relay and a block-producer yet) I am wondering, are relays running the same code as cardano nodes? If not, do relays have to be rebuilt as often as nodes.

2

Hello,

I’m wondering how often do you have to rebuild a cardano-node. And what are the signals/events that make you decide a rebuild of a cardano-node ?

Everytime when a new version is release, u need to update the nodes

Moreover, (I have not figured out the difference between a relay and a block-producer yet) I am wondering, are relays running the same code as cardano nodes? If not, do relays have to be rebuilt as often as nodes.

All of them are full nodes and will process transactions, but Block Producer is responsable with producing the blocks and relays to propagate them into the network… another important role for relays is to keep ur Producer “hiden” for the public network (security reason)

Cheers,

1 Like
3

And what is an acceptable delay once a new version is released ? Is it in hours, in days, in weeks ?
What is the consequence of not being up to date ? No block to produce, no reward? Can it be less constraining sometimes?

4

Hello,

And what is an acceptable delay once a new version is released ? Is it in hours, in days, in weeks ?

No one will force u to make the update imediately; u will do when u have time

What is the consequence of not being up to date ? No block to produce, no reward? Can it be less constraining sometimes?

It depends the update; can be an update for fixing something (bugs for example) or can be an update which will prepare the network to support/add new features (for example era transition or HardFork); u will have enough time to prepare the update; keeping ur nodes unupdated could affect the normal functionality of ur nodes and implicit no blocks/rewards (but again , depending of update)

Cheers,

2 Likes
5

if you are big pool with high saturation, you will have many blocks per epoch.
for that reason, if you want to upgrade your nodes, you should plan your process with High Availability and Redundancy.

small simple example i did develop in the past was the one below.
setup : 2 x Relays, 1x BP , 1x BP_Backup
all run Version_A, need to be upgraded to Version_B
steps to upgrade:

  1. all nodes run Version_A
  2. Relay_1 upgrade to Version_B
  3. BP_Backup upgrade to Version_B
  4. connect to Network BP_Backup, stop service on BP
  5. upgrade BP and Relay_2
  6. Start service Relay_2 with the Version_B
  7. BP → will be your new BP_Backup
2 Likes