CTE may stuck in register OOM action when got killed by global memory limit #43749
Labels
affects-5.1
This bug affects 5.1.x versions.
affects-5.2
This bug affects 5.2.x versions.
affects-5.3
This bug affects 5.3.x versions.
affects-5.4
This bug affects the 5.4.x(LTS) versions.
affects-6.0
affects-6.1
This bug affects the 6.1.x(LTS) versions.
affects-6.2
affects-6.3
affects-6.4
affects-6.5
This bug affects the 6.5.x(LTS) versions.
affects-6.6
affects-7.0
affects-7.1
This bug affects the 7.1.x(LTS) versions.
severity/major
sig/execution
SIG execution
type/bug
The issue is confirmed as a bug.
Bug Report
Please answer these questions before submitting your issue. Thanks!
1. Minimal reproduce step (Required)
If we got panic(e.g. oom cancel) in computeSeedPart,
e.resTbl
will fail to set err info, which makes other CTEExec that share the sameCTEExec.resTbl
register same memTracker again, which make dead loop of memTracker action rearrangementReproduce steps(We use failpoint to reproduce because it's hard to reproduce using SQL):
2. What did you expect to see? (Required)
sql runs with error
3. What did you see instead (Required)
sql stuck
4. What is your TiDB version? (Required)
master(79729c3)
The text was updated successfully, but these errors were encountered: