pingcap/tidb

br: wait till info schema loading finished before proceeding to the rest of the PiTR restore

Closed this issue · 1 comments

Bug Report

In a cluster with large number of tables but minimal actual data, during PiTR restore infoSchema can take minutes to load the restore change, and since actual data restore finished within seconds, restore process tries to use infoSchema for insertGCRow before its reloading finished, and it can cause problem of the below

[error="[domain:8027]Information schema is out of date: schema failed to update in 1 lease, please make sure TiDB can connect to TiKV"]

/report customer