RED-196 check trie hashes for near node down check #236
Conversation
PR Reviewer Guide 🔍
|
| let hashTrieReq: HashTrieReq = { | ||
| radixList: ['0'], | ||
| } | ||
| let closestNodes = stateManager.getClosestNodes(hash, 5, true) |
There was a problem hiding this comment.
How do we know if these 5 nodes will cover the radix 0 ?
|
|
||
| if (nearNodeResponse == null) { | ||
| /* prettier-ignore */ nestedCountersInstance.countEvent('p2p', 'isDownCheck-down-near-node-fail', 1) | ||
| return 'down' |
There was a problem hiding this comment.
Does it mean "near node" is down or "target node" is down? If we cannot get response from near node, isn't that we report it as "lost" and try another near node?
| } | ||
| //compare the hashes from the near node and the target node | ||
| for (let nodeHash of targetNodeResponse.nodeHashes) { | ||
| if (nodeHashesMap.get(nodeHash.radix) !== nodeHash.hash) { |
There was a problem hiding this comment.
What if the target node is up and responsive but has out of sync data? In that case, we should not mark it as down as the node may repair the radix in next cycle.
| if (config.p2p.useNearNodeForDownCheck) { | ||
| //using the 'get_trie_hashes' route to check if the node is up. | ||
| let obj = { counter: currentCycle, checker: Self.id, target: node.id, timestamp: shardusGetTime() } | ||
| let hash = crypto.hash(obj) |
There was a problem hiding this comment.
why is this hash used instead of a radix covered by target node?
No description provided.