log-backup: failed to restore index using PiTR

[joccau]

Bug Report

Please answer these questions before submitting your issue. Thanks!

The detail issue on https://pingcap.feishu.cn/wiki/wikcnPeLF5XEf1Ls04e6vpWdIMd

1. TiDB version
   6.2.0-alpha TiDB cluster and br

2. Reproduce steps.

• Start log backup
• Do full backup
• Start workload with 1000 tables for 10 min
  sysbench --db-driver=mysql --mysql-host=upstream-tidb.brie-acceptance-pitr-robust-tps-1083099-1-452 --mysql-port=4000 --mysql-user=root --mysql-db=sbtest1000 --tables=1000 --table-size=10000 --create_secondary=off --time=60 --threads=20 oltp_insert prepare sysbench --db-driver=mysql --mysql-host=upstream-tidb.brie-acceptance-pitr-robust-tps-1083099-1-452 --mysql-port=4000 --mysql-user=root --mysql-db=sbtest1000 --tables=1000 --table-size=10000 --create_secondary=off --time=600 --threads=20 oltp_insert run
• Wait checkpoint move forward
  `
  ● Total 1 Tasks.

#1 <
name: task1
status: ● NORMAL
start: 2022-08-13 02:28:36.516 +0000
end: 2090-11-18 14:07:45.624 +0000
storage: s3://tmp/br-log-backup84ed324a-894c-4f13-84fa-44675095787d/Log
speed(est.): 0.00 ops/s
checkpoint[global]: 2022-08-13 02:43:07.216 +0000; gap=4m59s
[2022/08/13 10:48:06.135 +08:00] [INFO] [br.go:518] ["Current log backup checkpoint: 2022-08-13 02:43:07 +0000 UTC, expect it is larger than: 2022-08-13 02:41:05.816 +0000 UTC"]
`

• Do a PiTR restore, and check data consistency
  /br restore point "--storage" "s3://tmp/br-log-backup84ed324a-894c-4f13-84fa-44675095787d/Log" "--pd" "http://downstream-pd.brie-acceptance-pitr-robust-tps-1083099-1-452:2379" "--full-backup-storage" "s3://tmp/br-log-backup84ed324a-894c-4f13-84fa-44675095787d/Full"

You will find the index on table in down steam cluster is disappeared.

[joccau]

/assign