Regression in neard testnet behavior

[frol]

Describe the bug

There is a regression in the default behavior of testnet neard subcommand between 1.19.2 and 1.20.0, which caused nayduck tests to fail: #4464 (comment). Initializing a new cluster of localnet nodes for testing now has the following values:

   "num_block_producer_seats_per_shard": [
-    4
+    1,
+    1,
+    1,
+    1
   ],
   "avg_hidden_validator_seats_per_shard": [
+    0,
+    0,
+    0,
     0
   ],

Full genesis.json diffs:

--- /tmp/near-1.19.2/node0/genesis.json	2021-07-26 15:35:35.583899833 +0200
+++ /tmp/near-1.20.0/node0/genesis.json	2021-07-26 15:35:45.384005185 +0200
@@ -1,13 +1,19 @@
 {
-  "protocol_version": 45,
-  "genesis_time": "2021-07-26T13:35:35.595752612Z",
-  "chain_id": "test-chain-44x3g",
+  "protocol_version": 46,
+  "genesis_time": "2021-07-26T13:35:45.392565119Z",
+  "chain_id": "test-chain-kBWkp",
   "genesis_height": 0,
   "num_block_producer_seats": 4,
   "num_block_producer_seats_per_shard": [
-    4
+    1,
+    1,
+    1,
+    1
   ],
   "avg_hidden_validator_seats_per_shard": [
+    0,
+    0,
+    0,
     0
   ],
   "dynamic_resharding": false,
@@ -154,6 +160,9 @@
         "keccak256_byte": 21471105,
         "keccak512_base": 5811388236,
         "keccak512_byte": 36649701,
+        "ripemd160_base": 853675086,
+        "ripemd160_block": 680107584,
+        "ecrecover_base": 3365369625000,
         "log_base": 3543313050,
         "log_byte": 13198791,
         "storage_write_base": 64196736000,
@@ -203,6 +212,7 @@
         "max_arguments_length": 4194304,
         "max_length_returned_data": 4194304,
         "max_contract_size": 4194304,
+        "max_transaction_size": 4194304,
         "max_length_storage_key": 4194304,
         "max_length_storage_value": 4194304,
         "max_promises_per_function_call_action": 1024,
@@ -258,7 +268,8 @@
           "amount": "950000000000000000000000000000000",
           "locked": "50000000000000000000000000000000",
           "code_hash": "11111111111111111111111111111111",
-          "storage_usage": 0
+          "storage_usage": 0,
+          "version": "V1"
         }
       }
     },
@@ -279,7 +290,8 @@
           "amount": "950000000000000000000000000000000",
           "locked": "50000000000000000000000000000000",
           "code_hash": "11111111111111111111111111111111",
-          "storage_usage": 0
+          "storage_usage": 0,
+          "version": "V1"
         }
       }
     },
@@ -300,7 +312,8 @@
           "amount": "950000000000000000000000000000000",
           "locked": "50000000000000000000000000000000",
           "code_hash": "11111111111111111111111111111111",
-          "storage_usage": 0
+          "storage_usage": 0,
+          "version": "V1"
         }
       }
     },
@@ -321,7 +334,8 @@
           "amount": "950000000000000000000000000000000",
           "locked": "50000000000000000000000000000000",
           "code_hash": "11111111111111111111111111111111",
-          "storage_usage": 0
+          "storage_usage": 0,
+          "version": "V1"
         }
       }
     },
@@ -342,7 +356,8 @@
           "amount": "1000000000000000000000000000000000",
           "locked": "0",
           "code_hash": "11111111111111111111111111111111",
-          "storage_usage": 0
+          "storage_usage": 0,
+          "version": "V1"
         }
       }
     },

To Reproduce

./near-1.19.2 --home /tmp/near-1.19.2 testnet

./near-1.20.0 --home /tmp/near-1.20.0 testnet

Expected behavior

I am not sure if that is expected, and what exactly the cause for that.

Version (please complete the following information):

• nearcore commit/branch: 1.19.x vs 1.20.x
• localnet

[bowenwang1996]

This looks like an unexpected change in the testnet command of neard

[frol]

@miraclx Can you, please, identify what introduced the change to num_block_producer_seats_per_shard and avg_hidden_validator_seats_per_shard values? If that is trivial to fix, feel free to submit a PR.

[mzhangmzz]

It's caused by https://github.com/near/nearcore/pull/4356/files

Before the change, default value of num_shards is 1 (https://github.com/near/nearcore/pull/4356/files#diff-e5cc0a30d85b6a9ab9b54636d706b348e4e25c560359c74945ff0bee48649e2bL186) while it was 4 after (https://github.com/near/nearcore/pull/4356/files#diff-2a66709b15f473bc8946b053ef86e3fc266329df2631afd7758744cfa3133137R245)