Hi team,

We reproduced this phenomenon on our side again.
Have a look at following operations, and its log, then look at the point
we mentioned it with "<-- *** This should be slave2 ***".

-bash-3.2$ LANG=C date ; psql -h 10.1.1.187 -p 9999 -c "show pool_nodes"
Thu Apr 25 21:03:54 JST 2013
 node_id | hostname | port | status | lb_weight |  role
---------+----------+------+--------+-----------+---------
 0       | slave1   | 5444 | 2      | 0.333333  | primary
 1       | slave2   | 5444 | 2      | 0.333333  | standby
 2       | slave3   | 5444 | 2      | 0.333333  | standby
(3 行)

execute following command on slave1.

-bash-3.2$ pg_ctl -D /opt/PostgresPlus/9.2AS/data -m immediate stop


We added "bash -x" for SRHS_follow_master.sh to see its details.

We got log files like below.

/tmp/pgpool.log
---
+ PG_CTL=/opt/PostgresPlus/9.2AS/bin/pg_ctl
+ FAILED_NODE_ID=0
+ FAILED_NODE_NAME=slave1
+ FAILED_NODE_PORT=5444
+ FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
+ NEW_MASTER_ID=1
+ OLD_MASTER_ID=0
+ NEW_MASTER_NAME=slave2
+ OLD_PRIMARY_ID=0
+ NEW_MASTER_DATA=/opt/PostgresPlus/9.2AS/data
+ TRIGGER=/tmp/trigger_file1
+ IDENTITY_FILE=/root/.ssh/id_rsa
+ '[' 0 = 0 ']'
++ date
+ echo $'2013\345\271\264' $'4\346\234\210' $'25\346\227\245' $'\346\234\250\346\233\234\346\227\245' 21:06:25 JST
+ echo FAILED_NODE_ID =0
+ echo FAILED_NODE_NAME=slave1
+ echo FAILED_NODE_PORT=5444
+ echo FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
+ echo NEW_MASTER_ID =1
+ echo OLD_MASTER_ID =0
+ echo NEW_MASTER_NAME =slave2
+ echo OLD_PRIMARY_ID =0
+ echo NEW_MASTER_DATA =/opt/PostgresPlus/9.2AS/data
+ ssh -T enterprisedb@slave2 /opt/PostgresPlus/9.2AS/bin/pg_ctl promote -D /opt/PostgresPlus/9.2AS/data
サーバを昇進中です。
+ echo 0
+ exit 0
+ LOGFILE=/var/log/pgpool/follow_master.log
+ PCP_ATTACH_NODE=/opt/PostgresPlus/9.2AS/bin/pcp_attach_node
+ FOLLOW_CMD=/opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh
+ PG_CTL=/opt/PostgresPlus/9.2AS/bin/pg_ctl
+ IDENTITY_FILE=/root/.ssh/id_rsa
+ TIMEOUT=10
+ HOST=localhost
+ PORT=9898
+ USERID=enterprisedb
+ PASSWD=redhat
+ FAILED_NODE_ID=0
+ FAILED_NODE_NAME=slave1
+ FAILED_NODE_PORT=5444
+ FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
+ NEW_MASTER_ID=1
+ OLD_MASTER_ID=0
+ NEW_MASTER_NAME=slave2
+ OLD_PRIMARY_ID=0
+ NEW_MASTER_DATA=/opt/PostgresPlus/9.2AS/data
++ date
+ echo $'2013\345\271\264' $'4\346\234\210' $'25\346\227\245' $'\346\234\250\346\233\234\346\227\245' 21:06:26 JST
+ echo FAILED_NODE_ID =0
+ echo FAILED_NODE_NAME=slave1
+ echo FAILED_NODE_PORT=5444
+ echo FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
+ echo NEW_MASTER_ID =1
+ echo OLD_MASTER_ID =0
+ echo NEW_MASTER_NAME =slave2
+ echo OLD_PRIMARY_ID =0
+ echo NEW_MASTER_DATA =/opt/PostgresPlus/9.2AS/data
+ NEW_MASTER_HOST=slave2
+ SLAVE_HOST=slave1
+ SLAVE_BASEDIR=/opt/PostgresPlus/9.2AS/data
+ NODE_ID=0
++ date '+%Y/%m/%d %H:%M:%S'
+ echo 2013/04/25 21:06:26 START '(/opt/PostgresPlus/9.2AS/data/SRHS_follow_master.sh' 0 slave1 5444 /opt/PostgresPlus/9.2AS/data 1 0 slave2 0 '/opt/PostgresPlus/9.2AS/data)'
+ ssh -T enterprisedb@slave2 /opt/PostgresPlus/9.2AS/bin/pg_ctl promote -D /opt/PostgresPlus/9.2AS/data
pg_ctl: サーバを昇進できません。サーバはスタンバイモードではありません。
++ ssh -i /root/.ssh/id_rsa enterprisedb@slave1 /opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh slave2 slave1 /opt/PostgresPlus/9.2AS/data
+ MSG='/opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh: line 22: /var/log/pgpool/follow_master.log: そのようなファイルやディレクトリはありません
touch: cannot touch `SRHS_follow_sub.sh.lock'\'': 許可がありません
/opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh: line 65: /var/log/pgpool/follow_master.log: そのようなファイルやディ レクトリはありません
サーバの起動完了を待っています....完了
サーバ起動完了
/opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh: line 75: /var/log/pgpool/follow_master.log: そのようなファイルやディ レクトリはありません'
+ RC=0
+ '[' 0 '!=' 0 ']'
++ date '+%Y/%m/%d %H:%M:%S'
+ echo 2013/04/25 21:06:28 PostgreSQL 'Restarted(slave1)'
+ /opt/PostgresPlus/9.2AS/bin/pcp_attach_node 10 localhost 9898 enterprisedb redhat 0
+ RC=0
+ '[' 0 '!=' 0 ']'
++ date '+%Y/%m/%d %H:%M:%S'
+ echo 2013/04/25 21:06:28 END '(/opt/PostgresPlus/9.2AS/data/SRHS_follow_master.sh)'
+ exit 0
+ LOGFILE=/var/log/pgpool/follow_master.log
+ PCP_ATTACH_NODE=/opt/PostgresPlus/9.2AS/bin/pcp_attach_node
+ FOLLOW_CMD=/opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh
+ PG_CTL=/opt/PostgresPlus/9.2AS/bin/pg_ctl
+ IDENTITY_FILE=/root/.ssh/id_rsa
+ TIMEOUT=10
+ HOST=localhost
+ PORT=9898
+ USERID=enterprisedb
+ PASSWD=redhat
+ FAILED_NODE_ID=2
+ FAILED_NODE_NAME=slave3
+ FAILED_NODE_PORT=5444
+ FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
+ NEW_MASTER_ID=1
+ OLD_MASTER_ID=0
+ NEW_MASTER_NAME=slave1
+ OLD_PRIMARY_ID=0
+ NEW_MASTER_DATA=/opt/PostgresPlus/9.2AS/data
++ date
+ echo $'2013\345\271\264' $'4\346\234\210' $'25\346\227\245' $'\346\234\250\346\233\234\346\227\245' 21:06:28 JST
+ echo FAILED_NODE_ID =2
+ echo FAILED_NODE_NAME=slave3
+ echo FAILED_NODE_PORT=5444
+ echo FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
+ echo NEW_MASTER_ID =1
+ echo OLD_MASTER_ID =0
+ echo NEW_MASTER_NAME =slave1
+ echo OLD_PRIMARY_ID =0
+ echo NEW_MASTER_DATA =/opt/PostgresPlus/9.2AS/data
+ NEW_MASTER_HOST=slave1
+ SLAVE_HOST=slave3
+ SLAVE_BASEDIR=/opt/PostgresPlus/9.2AS/data
+ NODE_ID=2
++ date '+%Y/%m/%d %H:%M:%S'
+ echo 2013/04/25 21:06:28 START '(/opt/PostgresPlus/9.2AS/data/SRHS_follow_master.sh' 2 slave3 5444 /opt/PostgresPlus/9.2AS/data 1 0 slave1 0 '/opt/PostgresPlus/9.2AS/data)'
+ ssh -T enterprisedb@slave1 /opt/PostgresPlus/9.2AS/bin/pg_ctl promote -D /opt/PostgresPlus/9.2AS/data
サーバを昇進中です。
++ ssh -i /root/.ssh/id_rsa enterprisedb@slave3 /opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh slave1 slave3 /opt/PostgresPlus/9.2AS/data
+ MSG='/opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh: line 22: /var/log/pgpool/follow_master.log: そのようなファイルやディレクトリはありません
touch: cannot touch `SRHS_follow_sub.sh.lock'\'': 許可がありません
/opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh: line 65: /var/log/pgpool/follow_master.log: そのようなファイルやディ レクトリはありません
サーバの起動完了を待っています....完了
サーバ起動完了
/opt/PostgresPlus/9.2AS/data/SRHS_follow_sub.sh: line 75: /var/log/pgpool/follow_master.log: そのようなファイルやディ レクトリはありません'
+ RC=0
+ '[' 0 '!=' 0 ']'
++ date '+%Y/%m/%d %H:%M:%S'
+ echo 2013/04/25 21:06:30 PostgreSQL 'Restarted(slave3)'
+ /opt/PostgresPlus/9.2AS/bin/pcp_attach_node 10 localhost 9898 enterprisedb redhat 2
+ RC=0
+ '[' 0 '!=' 0 ']'
++ date '+%Y/%m/%d %H:%M:%S'
+ echo 2013/04/25 21:06:30 END '(/opt/PostgresPlus/9.2AS/data/SRHS_follow_master.sh)'
+ exit 0
---

/var/log/messages
---
Apr 25 21:06:25 master1 pgpool[6647]: connect_inet_domain_socket: connect() failed: Connection refused
Apr 25 21:06:25 master1 pgpool[6647]: make_persistent_db_connection: connection to slave1(5444) failed
Apr 25 21:06:25 master1 pgpool[6647]: connect_inet_domain_socket: connect() failed: Connection refused
Apr 25 21:06:25 master1 pgpool[6647]: make_persistent_db_connection: connection to slave1(5444) failed
Apr 25 21:06:25 master1 pgpool[6647]: health check failed. 0 th host slave1 at port 5444 is down
Apr 25 21:06:25 master1 pgpool[6647]: set 0 th backend down status
Apr 25 21:06:25 master1 pgpool[6647]: starting degeneration. shutdown host slave1(5444)
Apr 25 21:06:25 master1 pgpool[6647]: Restart all children
Apr 25 21:06:25 master1 pgpool[6647]: execute command: /opt/PostgresPlus/9.2AS/data/SRHS_failover.sh 0 "slave1" 5444 /opt/PostgresPlus/9.2AS/data 1 0 "slave2" 0 /opt/PostgresPlus/9.2AS/data
Apr 25 21:06:25 master1 pgpool[6647]: find_primary_node_repeatedly: waiting for finding a primary node
Apr 25 21:06:26 master1 pgpool[6647]: find_primary_node: primary node id is 1
Apr 25 21:06:26 master1 pgpool[6647]: starting follow degeneration. shutdown host slave1(5444)
Apr 25 21:06:26 master1 pgpool[6647]: starting follow degeneration. shutdown host slave3(5444)
Apr 25 21:06:26 master1 pgpool[6647]: failover: 2 follow backends have been degenerated
Apr 25 21:06:26 master1 pgpool[6745]: start triggering follow command.
Apr 25 21:06:26 master1 pgpool[6745]: execute command: /opt/PostgresPlus/9.2AS/data/SRHS_follow_master.sh 0 "slave1" 5444 /opt/PostgresPlus/9.2AS/data 1 0 "slave2" 0 /opt/PostgresPlus/9.2AS/data
Apr 25 21:06:26 master1 pgpool[6647]: failover: set new primary node: 1
Apr 25 21:06:26 master1 pgpool[6647]: failover: set new master node: 1
Apr 25 21:06:26 master1 pgpool[6681]: worker process received restart request
Apr 25 21:06:26 master1 pgpool[6647]: failover done. shutdown host slave1(5444)
Apr 25 21:06:27 master1 pgpool[6647]: worker child 6681 exits with status 256
Apr 25 21:06:27 master1 pgpool[6647]: fork a new worker child pid 6785
Apr 25 21:06:27 master1 pgpool[6680]: pcp child process received restart request
Apr 25 21:06:27 master1 pgpool[6647]: PCP child 6680 exits with status 256
Apr 25 21:06:27 master1 pgpool[6647]: fork a new PCP child pid 6786
Apr 25 21:06:28 master1 pgpool[6786]: send_failback_request: fail back 0 th node request from pid 6786
Apr 25 21:06:28 master1 pgpool[6647]: starting fail back. reconnect host slave1(5444)
Apr 25 21:06:28 master1 pgpool[6647]: Do not restart children because we are failbacking node id 0 hostslave1 port:5444 and we are in streaming replication mode
Apr 25 21:06:28 master1 pgpool[6647]: find_primary_node_repeatedly: waiting for finding a primary node
Apr 25 21:06:28 master1 pgpool[6745]: execute command: /opt/PostgresPlus/9.2AS/data/SRHS_follow_master.sh 2 "slave3" 5444 /opt/PostgresPlus/9.2AS/data 1 0 "slave1" 0 /opt/PostgresPlus/9.2AS/data
Apr 25 21:06:28 master1 pgpool[6647]: find_primary_node: primary node id is 1
Apr 25 21:06:28 master1 pgpool[6647]: failover: set new primary node: 1
Apr 25 21:06:28 master1 pgpool[6647]: failover: set new master node: 0
Apr 25 21:06:28 master1 pgpool[6647]: failback done. reconnect host slave1(5444)
Apr 25 21:06:28 master1 pgpool[6785]: worker process received restart request
Apr 25 21:06:29 master1 pgpool[6647]: worker child 6785 exits with status 256
Apr 25 21:06:29 master1 pgpool[6647]: fork a new worker child pid 6796
Apr 25 21:06:29 master1 pgpool[6786]: pcp child process received restart request
Apr 25 21:06:29 master1 pgpool[6647]: PCP child 6786 exits with status 256
Apr 25 21:06:29 master1 pgpool[6647]: fork a new PCP child pid 6797
Apr 25 21:06:30 master1 pgpool[6797]: send_failback_request: fail back 2 th node request from pid 6797
Apr 25 21:06:30 master1 pgpool[6647]: starting fail back. reconnect host slave3(5444)
Apr 25 21:06:30 master1 pgpool[6647]: Do not restart children because we are failbacking node id 2 hostslave3 port:5444 and we are in streaming replication mode
Apr 25 21:06:30 master1 pgpool[6647]: find_primary_node_repeatedly: waiting for finding a primary node
Apr 25 21:06:30 master1 pgpool[6647]: find_primary_node: primary node id is 0
Apr 25 21:06:30 master1 pgpool[6647]: failover: set new primary node: 0
Apr 25 21:06:30 master1 pgpool[6647]: failover: set new master node: 0
Apr 25 21:06:30 master1 pgpool[6647]: failback done. reconnect host slave3(5444)
Apr 25 21:06:30 master1 pgpool[6796]: worker process received restart request
Apr 25 21:06:31 master1 pgpool[6647]: worker child 6796 exits with status 256
Apr 25 21:06:31 master1 pgpool[6647]: fork a new worker child pid 6801
Apr 25 21:06:31 master1 pgpool[6797]: pcp child process received restart request
Apr 25 21:06:31 master1 pgpool[6647]: PCP child 6797 exits with status 256
Apr 25 21:06:31 master1 pgpool[6647]: fork a new PCP child pid 6802
---


-bash-3.2$ cat /tmp/hoge.hogege
2013年 4月 25日 木曜日 21:06:25 JST
FAILED_NODE_ID =0
FAILED_NODE_NAME=slave1
FAILED_NODE_PORT=5444
FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
NEW_MASTER_ID =1
OLD_MASTER_ID =0
NEW_MASTER_NAME =slave2
OLD_PRIMARY_ID =0
NEW_MASTER_DATA =/opt/PostgresPlus/9.2AS/data
0
-bash-3.2$ cat /tmp/fuga.fugaga
2013年 4月 25日 木曜日 21:06:26 JST
FAILED_NODE_ID =0
FAILED_NODE_NAME=slave1
FAILED_NODE_PORT=5444
FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
NEW_MASTER_ID =1
OLD_MASTER_ID =0
NEW_MASTER_NAME =slave2
OLD_PRIMARY_ID =0
NEW_MASTER_DATA =/opt/PostgresPlus/9.2AS/data
2013年 4月 25日 木曜日 21:06:28 JST
FAILED_NODE_ID =2
FAILED_NODE_NAME=slave3
FAILED_NODE_PORT=5444
FAILED_NODE_DATA=/opt/PostgresPlus/9.2AS/data
NEW_MASTER_ID =1
OLD_MASTER_ID =0
NEW_MASTER_NAME =slave1  <-- *** This should be slave2 ***
OLD_PRIMARY_ID =0
NEW_MASTER_DATA =/opt/PostgresPlus/9.2AS/data
-bash-3.2$

Hope this will help.