fubinzh

What is the reason TiCDC is blocked? @hicqu > TiCDC is blocked with some reasons, and TiKV reports this error

CDC disk full issue, tracked by https://github.com/pingcap/tiflow/issues/10641

Close this issue, as https://github.com/pingcap/tiflow/issues/10641 fixed.

/severity critical

@tonyxuqqi Is it possible to resolve the lock automatically somehow? Though we can unlock it manually following https://github.com/pingcap/tiflow/issues/4516#issuecomment-1049540773, it is quite difficult for a user to itendify/work around such issue.

> @fubinzh Does the chaos test always running or just running for a while(10min or 20min) and then stop chaos test let the system to recover by itself? I think...

> The TiDB is killed during the test and therefore remaining locks are uncleared until the GC kicks in. The question is if GC is enabled or tikv_gc_life_time is changed...

Run scenario ha-tidb-random-kill -> recovery -> ha-tikv-random-to-tidb-random-network-partition-last-for-10m for several times today with 7.1.0, on a tidb cluster with 3 pd/tidb/tikv running tpcc workload (100 warehouse, qps ~25k), the issue is...

/remove-severity critical

/severity moderate