After Mainnet spork is complete, system doesn't connect to it automatically.

[Ronak-59]

This happened a while back around 9th February 2022 Flow Mainnet Spork. It was fine until then but as soon as spork started, it started giving following error (which is okay and intentional I believe) :

level=warning msg="Unable to connect to chain, pausing system" function=Listener.Start.goroutine package=chain_events

But long after the spork was complete, the errors were still coming till 11th February 2022 when we checked and decided to restart the service on AWS ECS which started normally and was able to connect to Mainnet nodes. But every now and then, we are seeing the above error in logs randomly at a frequency of like around 5 times a day.

Do we need to restart the service after each Mainnet spork?

[Ronak-59]

We didn't find this behavior on Testnet. It automatically connected to testnet nodes after testnet spork around 8th February 2022.

[latenssi]

Hi @Ronak-59! There have been some connectivity issues on mainnet. Have you been able to get over a testnet spork without issues?

[Ronak-59]

Hey @latenssi, apologies for the delay in response. Yes this wallet-api automatically connects to testnet nodes once the testnet spork is complete. But not for Mainnet spork.

Considering that on restarting this service, it successfully connected to Mainnet nodes immediately, I don't think it was connectivity issues on Mainnet nodes. It seemed more like this wallet service was stuck in an endless loop to log the error every minute until it was restarted.

[rpatidar-systango]

I had the same problem after 6th April 2022 Mainnet Spork completion. Even after restarting it was not able to connect to Chain and the service was continuously throwing level=warning msg="Unable to connect to chain, pausing system" function=Listener.Start.goroutine package=chain_events. From the local system, the chain was getting connected but not from my production server. Then I created a new DB and connect the service to the newly created DB and there it was connected. not sure what status was being managed in DB. Only one configuration I found in the system_settings table where it was managing paused_since value with some date and time. but even setting it to null it was automatically reset to a datetime.