Closed adv0r closed 3 years ago
Same symptoms after updating to 1.0.2. Fallback n APIs to both alchemy and infura was tested and working fine till after the update. Infura seem to work fine without any errors but there isnt seem to be any calls logged at the dashboard. Alchemy simply doesnt work , gets a 403 error
I'm unable to reproduce this on my end across multiple systems. @adv0r and @preddy25 could you please share some details about your hosts. E.g., operating system, CPU architecture (ARM, Intel, etc), if you compiled lighthouse locally or used cross-compiling.
@blacktemplar, just bringing your attention to this in case you've faced this before.
Can you confirm that v1.0.1
works fine but v1.0.2
does not? (Or are you upgrading from an even earlier version?)
Not able to reproduce this either with multiple syncs.
@blacktemplar, just bringing your attention to this in case you've faced this before.
So during my tests I also encountered some problems with infura nodes, running into timeouts. Both errors could possibly be caused by a timeout (the first if the timeout happens during the initial network_id
/chain_id
checks and the second if a timeout happens afterwards during getting deposit logs).
@adv0r you said you can see the requests + responses in the Alchemy dashboard, do you also see by chance timestamps, when the requests arrived / when the responses were sent? Furthermore, are those errors happening consistently, or just sometimes?
Have the same issue. Running lighthouse-v1.0.2-x86_64-unknown-linux-gnu-portable
2 things I have noticed:
--eth1-endpoint
when using multiple ones@adv0r you have a trailing slash in the endpoint URL
@b-m-f I have --eth1-endpoints
ExecStart=/usr/local/bin/lighthouse bn --network mainnet --staking --datadir /var/lib/lighthouse --metrics --eth1-endpoints http://192.168.1.12:8545,https://eth-mainnet.alchemyapi.io/v2/HM*******j6g,https://mainnet.infura.io/v3/16c****aef5f
The WARN appears only after I kill the local eht1 endpoint:
Nov 30 19:48:48.620 WARN Error connecting to eth1 node. Trying fallback ..., endpoint: http://192.168.1.12:8545, service: eth1_rpc
Looking at alchemy, I see requests coming in:
@pg3141 That looks correct though.
After killing the local node the switch-over should occur.
Or do the remote endpoints time out as well?
@b-m-f I was under the impression that those WARN messages mean the switch over did not occur, but judging by the fact that requests do come in to the fallback endpoints... I guess the switch over does occur.
I guess the switch over does occur.
That's correct, therefore I understand that Lighthouse is performing as expected.
@adv0r you have a trailing slash in the endpoint URL
I can reproduce this issue if I add the trailing slash, leading me to think that the root cause is a malformed --eth2-endpoints
flag and therefore not an issue with Lighthouse.
I'm going to close this since I believe all issues are addressed, feel free to reopen if that's not the case.
Special thanks for @b-m-f for their help here :)
@adv0r you said you can see the requests + responses in the Alchemy dashboard, do you also see by chance timestamps, when the requests arrived / when the responses were sent? Furthermore, are those errors happening consistently, or just sometimes?
Error happens consistently...
To reply your question,as I said on discord where you can also see a screenshot.
by checking alchemy logs I can see that in less than 1 minute from starting, lighthouse makes more than 40+ queries
to the method eth_getlogs
( All the calls are succesful btw, with 200 ms response) .
moreover all the queries for the same fromblock, toblock numbers.
Is this expected? Something is not right, with that many requests in 1 minute, I will have to pay for millions/request per month soon, cluttering their service
@adv0r you have a trailing slash in the endpoint URL
Doesn't change the result
Is this expected? Something is not right, with that many requests in 1 minute, I will have to pay for millions/request per month soon, cluttering their service
This is expected, when it boots it needs to sync all the deposit logs. After that it will cache them and only request new ones.
Description
After updating lighthouse can't connect to eth1 nodes. This has been confirmed by another user on discord. We are both running ubuntu and tried both
alchemyapi
andinfura.io
.No firewalls, no mistakes in the endpoint.
I can curl the endpoint with no issues.
Version
cargo 1.48.0 (65cbdd2dc 2020-10-14) Lighthouse v1.0.2-f7183098 BLS Library: blst
Present Behaviour
Command:
lighthouse bn --network mainnet --datadir /mnt/ssd/stake/ --eth1-endpoint https://mainnet.infura.io/v3/6****f/
Output:
Some other time the output it's slighly different:
I have no authentication activated on Infura.
I can curl it just fine
Replace all these symptoms with Alchemy, same problem
In alchemy dashboard I can see all the request arriving to the endpoint with no problem. Here is the raw request:
and here is a raw response
Steps to resolve