Attester slashing 5212

[dankrad]

See https://medalla.beaconcha.in/validator/5212
(no blocks proposed, so graffiti not known)

[djrtwo]

They made these additional deposits https://medalla.beaconcha.in/validators/eth1deposits?q=0x89088e8559cb48887e3f496eAeA6988502944f22

No other blocks proposed for those validators yet either

[djrtwo]

This user is "AlexSSD7#8774 in Discord" on discord and probably the same @AlexSSD7 from github.

@AlexSSD7 can you help us understand your setup and any details that might have led to this slashing?

[AlexSSD7]

It is a Prysmatic Labs' Goerli faucet address. Probably just copied the wrong one and haven't noticed. xD

[djrtwo]

Sorry, scratch the above.

It's discord user baconBeacon

[terencechain]

I just dm'ed baconBeacon to learn more

[terencechain]

I spoke with baconBeacon Here's the summary:

It was a teku client running with 1 validator. It did not appear to crash or hang. Commands used to launch:

./teku --network=medalla --eth1-endpoint=http://localhost:8545/ --validators-files=keystore-m_....json --validators-key-password-files=tekupassword.txt --rest-api-enabled=true --rest-api-docs-enabled=true --metrics-enabled --validators-graffiti="poap..."

Logs: teku.zip

It's worth noting that this slashed key was rescued over from a lodestar client. I'm confirming whether teku and lodestar were both running at the same time

[ajsutton]

The logs look completely normal around slot 688 where they were slashed. It's not near a restart point and looks like the chain is progressing forward as you'd expect.

[dankrad]

Who was running the lodestar? Can we check whether it was still running at that point?

[djrtwo]

djrtwoToday at 10:07 AM
Might you be running the same keys on multiple clients? Or are you sure the keys are fully isolated?

baconBeaconToday at 1:56 PM
I don't believe so. I checked the imported keystore folders. They are different validator public keys.