BUG #14789: In Cascading Replication，the "wal_level" is setted to "archive",the operation becomes abnomal

The following bug has been logged on the website:

Bug reference: 14789
Logged by: pu qun
Email address: 691525127(at)qq(dot)com
PostgreSQL version: 9.2.19
Operating system: Red Hat Enterprise Linux Server release 6.0 (Santi
Description:

In
"https://wiki.postgresql.org/wiki/What%27s_new_in_PostgreSQL_9.2#Replication_improvements",
We can see the following description about "Cascading Replication".

"We set it up for replication:

postgresql.conf:
wal_level=hot_standby #(could be archive too)
max_wal_senders=5
hot_standby=on"

The parameter "wal_level" can be set to "hot_standby" or "archive" by the
above-mentioned.

However, when "wal_level" is set to archive, the cascade replication cannot
be done.

The sample is as follows.
1. First, create a primary server.
The setting of postgresql.conf is changed as follows:

wal_level = archive
max_wal_senders = 5
listen_addresses = '*'
hot_standby = off #（When "wal_level" is set to "archive", it is
necessary to specify "off" for "hot_standby". ）

2. Using pg_basebackup to create the first standby server.

3. After starting the standby 1,using pg_basebackup to create Standby 2.
Standby 2 is copied from standby 1.
It fails in pg_basebackup, and standby 2 cannot be generated with the
cascade replication.

The error is as follows:
[postgres92(at)host-193-160-38-95 inst]$ pg_basebackup -F p -D
/file1/standby2/inst/ -h 193.160.38.95 -p 55224 -U postgres92
FATAL: the database system is starting up
pg_basebackup: could not connect to server: FATAL: the database system is
starting up

4. The above issue is not generated when "wal_level" is set to
"hot_standby",not set to "archive".