Request for Information about Disputes #GDR-51

Governance & GIPs
4

Hi all, Hau from Pinax here :waving_hand: . Just sharing my perspective.

Based on input from Fisherman and discussion in #47, P2P indexer has already resolved the POI issue for subgraph QmdDSA73QkFAvRZqtRoHgLvxonuSZTuJhoWkmBtEdVuTmz at block 75,174,859.

Disputes #50, #51, #52, #53 and #54 appear to be new issues at later blocks, and should be investigated further to determine the root cause of the discrepancies.

From experience, subgraphs can still exhibit discrepancies at later blocks despite rewinding, clearing block and call caches, or performing a full resync.

It’s also worth noting that several of these discrepancies may occur at the same block height across multiple subgraphs. This pattern may indicate a shared upstream data inconsistency (e.g. RPC or Firehose input), rather than isolated subgraph or indexer-specific issues.

We are currently performing off-chain syncing for the following five subgraphs related to disputes #50–#54:

  • QmY1133fWsE94ZDtgJyTYjzxwZNdCJKXX3j6gVs4mxNNUb

  • QmdDSA73QkFAvRZqtRoHgLvxonuSZTuJhoWkmBtEdVuTmz

  • QmZyoRwhXKY2o9Yxqur5hNufKPPSwhZANtf4QQaiqC7Qms

  • QmNmddMQAaLL1zpwpA69tHjCom4udHcFJ5a9PZPxU5Tkew

  • QmTbmXhUr67gAt5eLy5zUo7dRqATG8S8QgPEcdVhmenRNU

This process may take approximately 2–3 days at best to complete verification. We’re available to assist with any Firehose-related questions if needed.

In our view, based on past disputes, indexers who used the indexer-agent as intended to close allocations should not be considered disputable, provided the subgraph was fully synced or only encountered deterministic errors.

Conversely, indexers who manually submitted POIs via direct contract calls without achieving full sync, or who bypassed non-deterministic errors, may be subject to slashing.

P2P indexer also mentioned :

1 Like