Storage space is not released

[GhostBoyBoy]

I tested it through pulsar-perf, the command:
./pulsar-perf consume -t 1 -n 1 -q 1000 -ns 1 -st Shared --acks-delay-millis 0 persistent://prop/nsp/my-topic
./pulsar-perf produce -t 1 -n 1 -r 1000000 persistent://prop/nsp/my-topic

Modified configuration：
1.This configuration is set by pulsar-admin.
[root@ecs-pulsar-broker-0003 bin]# ./pulsar-admin namespaces get-retention prop/nsp
{
"retentionTimeInMinutes" : 1,
"retentionSizeInMB" : 1
}

2.This configuration modify broker.conf and then restart the broker.
managedLedgerOffloadDeletionLagMs=1000

3.Partition is pulsar-admin, divided into 100

Others are the default configuration.

Problem：
1.The backlog is 0, why is the storage space not released?
2.How can I set up to free up space immediately after consumption of messages?

After I rolled and restarted the broker, it was still like this.

[GhostBoyBoy]

At that time, it was not released within a few minutes. I don’t know how long it took to release. I want to know the specific reason so that I can control it.

[GhostBoyBoy]

I found the reason, ledger did not switch.

Why is the entry more than 50000 and has not been switched?

[GhostBoyBoy]

@sijie

[GhostBoyBoy]

@sijie

It has been resolved that the managedLedgerMinLedgerRolloverTimeMinutes configuration causes the ledger to not switch and the space is not released. This configuration is very dangerous. Is the default value considered to be changed to 0.