Error in MPPTask.cpp:636 reported in TiFlash logs

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: tiflash日志中报MPPTask.cpp:636的错误

| username: Running

TiDB Usage Environment: Production Environment
TiDB Version: 7.5
The tiflash log occasionally reports an error at MPPTask.cpp:636. I want to know if it is caused by a specific SQL execution error or some internal mechanism.

| username: wangccsy | Original post link

Isn’t TiDB written in GO? Why is there CPP as well?

| username: Running | Original post link

You need to ask the official team about this. I want to know the truth behind the issue.

| username: onlyacat | Original post link

TiFlash is written in C++.

| username: tidb狂热爱好者 | Original post link

Take a look at the memory monitoring.

| username: Billmay表妹 | Original post link

Take a look at this.

| username: tidb狂热爱好者 | Original post link

TiFlash is a C++ software originally based on ClickHouse.

| username: Running | Original post link

This is not an issue, no error keyword “Different aggregation mode detected” was found; the current error is reported at MPPTask.cpp:636.

| username: WalterWj | Original post link

You can ask the R&D team for help in the feedback section.

| username: wzf0072 | Original post link

Check in Grafana to see if the TiFlash or TiDB Server nodes have restarted.

| username: dba远航 | Original post link

After looking at it for a long time, I still don’t understand it. The error report is too vague.

| username: 有猫万事足 | Original post link

Take a look at the instance monitoring of TiFlash. The inexplicable errors in MPP are mostly due to a certain TiFlash instance running out of memory and encountering an OOM (Out of Memory) issue during MPP computation. As a result, it suddenly loses contact with other TiFlash instances that are also performing MPP computations together.

| username: 有猫万事足 | Original post link

tidb—go
tikv—rust + rocksdb—c++
tiflash----c++