Sql-server – Database Log Reuse Wait shows AVAILABILITY_REPLICA but database is no longer part of an availability group

You can backup the transaction log from either of the replicas. Doing the transaction log backup on either the primary replica or the secondary replica will mark both replicas' transaction logs as reusable (provided no other stoppers are in place like active transactions, etc.).

To do a test, in a non-production environment, setup an availability group just as you have it in your production system (asynchronous commit to the secondary replica).

In my test environment I have a test database, TestBackupDatabase, and I bloated it with logged transactions through a dummy table:

use TestBackupDatabase;
go

create table dbo.TestTable
(
    id int identity(1, 1) not null,
    some_int int not null
        default 1
);
go

insert into dbo.TestTable
default values;
go 1000

Now when I do a transaction log backup on my primary, using DBCC SQLPERF(LOGSPACE) I see on both transaction logs (primary and secondary) that space used has dropped due to log truncation. Bloating the transaction log back up with the same test on the primary:

insert into dbo.TestTable
default values;
go 1000

I now do a transaction log backup on the secondary async replica. Running DBCC SQLPERF(LOGSPACE) again on each replica I see the same behavior: transaction log reuse.

BOL Reference: Active Secondaries: Backup on Secondary Replicas (AlwaysOn Availability Groups)

In AGs writes can only occur on the primary. Shrink operations are writes. Therefore you must do the shrink on the primary. Note that the shrink may not shrink as much as you expect, your test on the restored DB had probably leveraged simple recovery model. Read How to shrink the SQL Server log for more info.

Do not shrink to 160MB. Determine why did the log grow to 121Gb so it does not repeat (you have a suspicion, would be nice to confirm if possible). Size the log to a size appropriate for your operational needs. Log growth is a serious problem, it cannot use instant file initialization and all your database activity will freeze while the log grows and is being 0-initialized. Users and apps hate it when it occurs. If you understand the impact and your users are OK, you can shrink once to a small amount (160MB is probably too small though) and let it grow until it stabilizes.