Backup and Availability

Node
1

Hi support,

do you have any recommendations for backup and availability for node management on mainnet.

  • how to backup node so that when we recover from a big failure, then we don’t have to rebuild the whole chain and catch up? Any production setup recommendations on setup and backup?

  • Availability: how do we address availability? an active-passive VM setup for the node? Or will an active-active work for node grpc? Any production recommendation minimum setup?

2

Hi,

  • the way to do that is to back up the node’s data directory. The important thing to back up is the database-v4 directory inside it (and the back up is to just make a copy of it). You should stop the node before taking a backup of it to avoid issues.

  • I am not sure what you mean by active-active. Would you put a load balancer or proxy in front of it? With the load balancer you have to be prepared for data incosistencies in the sense that the two nodes do not have exactly the same state at all times (i.e., they will get blocks/transactions at slightly different times). We are in fact using this for running some of our services (an Amazon application load balancer) however the clients must be adapted to support load balancing cookies.

I refer to our documentation for minimum setup. However I will add that if you want to make use of parallelism for queries make sure to set the number of threads appropriately using

      - CONCORDIUM_NODE_RUNTIME_HASKELL_RTS_FLAGS=-N4,-I0

for example use -N4 to allow 4 threads.

1 Like
3

Thanks, yes load balancing is something we consider.