Pyth Daemon - "not ready to publish - check rpc / pyth_tx connection" on RPC reconnect

[opaolini]

pythd version - 2.9.1
command: home/pyth/pyth-client/build/pythd -r nginx:7899:7900 -t pyth-tx-server-mainnet:8898 -k $KDIR -w /home/pyth/pyth-client/dashboard"

When pythd reconnects to the Solana RPC we see either of the following logs:

[2021-11-29T14:37:24.923149Z 11 INF rpc_connected                           ]
[2021-11-29T14:37:24.896228Z 11 ERR rpc_websocket_reset                     ] error=fail to read [11 Resource temporarily unavailable],host=nginx:7899:7900,port=7899

[2021-11-29T05:09:28.396393Z 11 ERR rpc_http_reset                          ] error=fail to read [11 Resource temporarily unavailable],host=nginx:7899:7900,port=7899
[2021-11-29T05:09:28.451638Z 11 INF rpc_connected    

After the reconnect happens, we are unable to post price_update because of the following JSON RPC error:

Error code: -32002 message: not ready to publish - check rpc / pyth_tx connection

We don't see any matching pythx errors. The only way for us to fix this is to restart the pythd process.

Ideally pythd would be able to handle those JSON RPC requests without having to restart it after a RPC reconnect.

[jayantk]

we are aware of this issue. we're working on a rewrite for pythd that will resolve this issue along with the many other issues with pythd.

[tompntn]

As of v2.10.0 this is no longer a problem - the JRPC requests will succeed after a reconnect, and even if the network is disconnected.