Exception occurred when synchronizing blocks

[jakamobiii]

Software Versions

OS : Linux
JVM : Oracle Corporation 1.8.0_161 amd64
Version : 4.7.7

Expected behavior

The node is running normally and no exception should be thrown.

Actual behavior

An SR is deployed and only connected to fullnode A. After fullnode A is started, it had been synchronizing blocks normally for a while, but later it threw an exception. Does it have any impact on the system，and whether it will affect block synchronization?

Frequency

Only appeared once.

[xxo1shine]

@jakamobiii According to the log information, getBlockChainHashesOnFork failed, which may be related to your special network.

When Fullnode goes offline, SR is isolated, and it will generate blocks by itself and fork from the main chain.

After FullNode is started, it will synchronize blocks from SR and the main network at the same time. It is possible that as the main network blocks are processed and solidified, the blocks below the solidified blocks will be removed from KhaosDatabase. When processing the synchronization message from SR, it is possible that the parent blocks of the two chains may be removed from KhaosDatabase, resulting in the failure of getBlockChainHashesOnFork.

The connection will be disconnected after the exception, and it will return to normal when the connection is established again. The impact on the system is minimal and will not affect the synchronization and broadcast of blocks.

[jakamobiii]

@xxo1shine OK, I got it, thanks. Is there any plan to fix this issue?

[xxo1shine]

@jakamobiii Yes, considering that it has little impact on the system, it is planned to be fixed in the next version.

[fyyhtx]

@xxo1shine Is it possible for the scenario you described to occur in the production environment of TRON? After all, the possibility of SR being isolated in the production environment is relatively low.

[xxo1shine]

@fyyhtx In a production environment, the number of connections to an SR node is generally more than one, so it is not easy to be isolated. Therefore, in a production environment, this problem generally does not exist.

[xxo1shine]

Milestone Update（2025-05-22）

• Current Status：in-progress
• Next Steps：
  • Full-chain testing expected to be completed in late June
  • Submit code for review in late June
• Responsible Persons：@xxo1shine