在 Greenplum 不小心把 IPv6 loopback 規則註解掉的話怎辦

繼上次 EDB 企業版鎖住 Superuser 登入不了之後，再一個怎麼辦系列第二彈～
這次輪到 Greenplum。

如下手工做了這個動作～用 vi 把 IPv6 localhost ::1 登入規則給註解掉（這邊用 sed 示意），然後又重新載入資料庫設定。。。

[gpadmin@gpdb6x ~]$ sed -e 's/\(.*::1.*\)/#\1/g' -i /data/master/gpseg-1/pg_hba.conf
[gpadmin@gpdb6x ~]$ grep '\(.*::1.*\)' /data/master/gpseg-1/pg_hba.conf 
#host     all         gpadmin         ::1/128       trust
[gpadmin@gpdb6x ~]$ 
[gpadmin@gpdb6x ~]$ gpstop -u
20200822:15:12:45:003894 gpstop:gpdb5x:gpadmin-[INFO]:-Starting gpstop with args: -u
20200822:15:12:45:003894 gpstop:gpdb5x:gpadmin-[INFO]:-Gathering information and validating the environment...
20200822:15:12:45:003894 gpstop:gpdb5x:gpadmin-[INFO]:-Obtaining Greenplum Master catalog information
20200822:15:12:45:003894 gpstop:gpdb5x:gpadmin-[INFO]:-Obtaining Segment details from master...
20200822:15:12:45:003894 gpstop:gpdb5x:gpadmin-[INFO]:-Greenplum Version: 'postgres (Greenplum Database) 6.10.0 build commit:cced3a78d1a0650ba21205438955d2df06644bc8'
20200822:15:12:45:003894 gpstop:gpdb5x:gpadmin-[INFO]:-Signalling all postmaster processes to reload
. 
[gpadmin@gpdb6x ~]$ 

再來就會發現。。。沒辦法重新載入或關閉整組 Greenplum 叢集惹。。。

就算把 pg_hba.conf 改回來也沒辦法～

[gpadmin@gpdb6x ~]$ sed -e 's/#\(.*::1.*\)/\1/g' -i /data/master/gpseg-1/pg_hba.conf
[gpadmin@gpdb6x ~]$ grep '\(.*::1.*\)' /data/master/gpseg-1/pg_hba.conf 
host     all         gpadmin         ::1/128       trust
[gpadmin@gpdb6x ~]$ 
[gpadmin@gpdb6x ~]$ gpstop -u
20200822:15:14:05:003913 gpstop:gpdb5x:gpadmin-[INFO]:-Starting gpstop with args: -u
20200822:15:14:05:003913 gpstop:gpdb5x:gpadmin-[INFO]:-Gathering information and validating the environment...
20200822:15:14:05:003913 gpstop:gpdb5x:gpadmin-[INFO]:-Obtaining Greenplum Master catalog information
20200822:15:14:05:003913 gpstop:gpdb5x:gpadmin-[INFO]:-Obtaining Segment details from master...
20200822:15:14:05:003913 gpstop:gpdb5x:gpadmin-[CRITICAL]:-gpstop failed. (Reason='FATAL:  no pg_hba.conf entry for host "::1", user "gpadmin", database "template1", SSL off') exiting...
[gpadmin@gpdb5x ~]$ 

由於 Greenplum 並不像 PostgreSQL 資料庫一樣只有一個資料庫，而是有一群節點。因此。。感覺現在處於關不掉資料庫惹。。。難道只能手動執行很傷的 kill 指令了ㄇ。。？

其實～還是有辦法的～～～只要想到 Greenplum 的骨幹還是 PostgreSQL 資料庫，就會發現在資料庫裡面還是能 reload～透過 pg_reload_conf() 函數！

因此，只要把 pg_hba.conf 參數檔改回來（像上面一樣），再登入資料庫 reload 就可以了～

[gpadmin@gpdb6x ~]$ psql
psql (9.4.24)
Type "help" for help.

postgres=# select pg_reload_conf();
 pg_reload_conf 
----------------
 t
(1 row)

postgres=# \q
[gpadmin@gpdb6x ~]$ gpstop -u
20200822:15:14:40:003928 gpstop:gpdb5x:gpadmin-[INFO]:-Starting gpstop with args: -u
20200822:15:14:40:003928 gpstop:gpdb5x:gpadmin-[INFO]:-Gathering information and validating the environment...
20200822:15:14:40:003928 gpstop:gpdb5x:gpadmin-[INFO]:-Obtaining Greenplum Master catalog information
20200822:15:14:40:003928 gpstop:gpdb5x:gpadmin-[INFO]:-Obtaining Segment details from master...
20200822:15:14:40:003928 gpstop:gpdb5x:gpadmin-[INFO]:-Greenplum Version: 'postgres (Greenplum Database) 6.10.0 build commit:cced3a78d1a0650ba21205438955d2df06644bc8'
20200822:15:14:40:003928 gpstop:gpdb5x:gpadmin-[INFO]:-Signalling all postmaster processes to reload
. 
[gpadmin@gpdb6x ~]$ 

自手殘中復活成功～