3

I have been running two nodes on different locations on my server. The "full" mainnet-snapshots node was running on an SSD and the "archive" mainnet node has been running on the HDD. At the beginning I observed how the HDD node was lagging and taking a loooong time to advance. The SSD node was running smoothly. Now two days ago the situation has inverted. If I run both nodes at the same time, one is always lagging. The bootstrap peers are different for the two nodes in the config files.

I really wanted to keep an "archive" node running on the HDD, while using the SSD node for baking. But this way it just makes my SSD node to lag 40min behind the chain and show "p2p.maintenance:Too few connections" errors.

Am I maybe doing something wrong? Or this is configuration just plainly not recommended and the principle should be "1 Machine, 1 Node"?

Improve this question
2
  • 1
    Can you elaborate on the setup? Are these different machines? Are you behind a NAT? Are both nodes running on the same port?
    – Arthur B
    May 4 '19 at 18:57
  • Hi Arthur, thanks for following up. I finally managed to get both nodes to sync and they have been working fine for 2 days now. Apparently the Mainnet-snapshots node got corrupted and I ended up having to export and import a new snapshot.
    – Javi Hache
    May 5 '19 at 21:45
1

I finally managed to get both nodes to sync and they have been working fine for two days now. Apparently the Mainnet-snapshots node got corrupted and I ended up having to export and import a new snapshot.

The nodes are in the same Ubuntu 18.04 machine and were using different ports for RPC and listen. The server is behind a router but the ports are open for both nodes. It would be similar to the "example setup" in this document: http://tezos.gitlab.io/mainnet/releases/april.html

I would recommend to follow the steps in that document one by one. I myself did it in a similar way, but the instructions are really clear. Please also take enough time between endorsements to avoid missing them like I did, due to the time it takes to do some of the tasks like exporting.

Improve this answer
2
  • the link is dead, could you please update ? thanks
    – Ezy ♦
    Jul 13 '19 at 11:05
  • Hi, it's irrelevant today, since Mainnet contains Snapshots already. You can basically use the example setup of the Documentation on Mainnet for running two nodes. I basically run one "full" node now and no "archive" node anymore.
    – Javi Hache
    Jul 16 '19 at 23:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.