Do you mean add standby config into hawq-site and restart? I have tried in the environment which has activate standby but seem failed:
After update all config: [gpadmin@test-1 ~]$ grep standby /usr/local/apache-hawq/etc/hawq-site.xml <name>hawq_standby_address_host</name> <description>The host name of hawq standby.</description> <name>hawq_standby_address_port</name> stop cluster and start : [gpadmin@test-1 ~]$ hawq start cluster -a 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-Prepare to do 'hawq start' 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-You can find log in: 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-/home/gpadmin/hawqAdminLogs/hawq_start_20181101.log 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-GPHOME is set to: 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-/usr/local/apache-hawq 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-Start hawq with args: ['start', 'cluster'] 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-Gathering information and validating the environment... 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-Start all the nodes in hawq cluster 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-Starting standby master '192.168.50.114' 20181101:10:03:11:006093 hawq_start:test-1:gpadmin-[INFO]:-Start standby master service 20181101:10:03:11:1080069 hawqstandbywatch.py:test-0:gpadmin-[INFO]:-Checking standby master status 20181101:10:03:11:1080069 hawqstandbywatch.py:test-0:gpadmin-[INFO]:-Monitoring logs 20181101:10:03:14:1080069 hawqstandbywatch.py:test-0:gpadmin-[INFO]:-checking if syncmaster is running 20181101:10:03:15:1080069 hawqstandbywatch.py:test-0:gpadmin-[INFO]:-syncmaster appears ok, pid 1080054 20181101:10:03:15:006093 hawq_start:test-1:gpadmin-[INFO]:-Standby master started successfully 20181101:10:03:15:006093 hawq_start:test-1:gpadmin-[INFO]:-Starting master node '192.168.50.115' 20181101:10:03:15:006093 hawq_start:test-1:gpadmin-[INFO]:-Start master service 20181101:10:03:16:006093 hawq_start:test-1:gpadmin-[INFO]:-Checking if standby is synced with master 20181101:10:03:16:006093 hawq_start:test-1:gpadmin-[INFO]:-Waiting standby to get synced for 3 seconds.. 20181101:10:04:17:006093 hawq_start:test-1:gpadmin-[INFO]:-Waiting standby to get synced for 3 seconds... 20181101:10:04:20:006093 hawq_start:test-1:gpadmin-[WARNING]:-Standby is not synced after 60 seconds 20181101:10:04:20:006093 hawq_start:test-1:gpadmin-[WARNING]:-Use 'hawq init standby -n' to do force sync 20181101:10:04:20:006093 hawq_start:test-1:gpadmin-[INFO]:-Master started successfully 20181101:10:04:20:006093 hawq_start:test-1:gpadmin-[INFO]:-Start all the segments in hawq cluster 20181101:10:04:20:006093 hawq_start:test-1:gpadmin-[INFO]:-Start segments in list: ['192.168.50.114', '192.168.50.115'] 20181101:10:04:20:006093 hawq_start:test-1:gpadmin-[INFO]:-Start segment service 20181101:10:04:20:006093 hawq_start:test-1:gpadmin-[INFO]:-Total segment number is: 2 ... 20181101:10:04:23:006093 hawq_start:test-1:gpadmin-[INFO]:-2 of 2 segments start successfully 20181101:10:04:23:006093 hawq_start:test-1:gpadmin-[INFO]:-Segments started successfully 20181101:10:04:23:006093 hawq_start:test-1:gpadmin-[INFO]:-HAWQ cluster started successfully [gpadmin@test-1 ~]$ hawq state 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- HAWQ instance status summary 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:------------------------------------------------------ 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Master instance = Active 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Master standby = None 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Standby master state = 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Total segment instance count from config file = 2 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:------------------------------------------------------ 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Segment Status 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:------------------------------------------------------ 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Total segments count from catalog = 2 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Total segment valid (at master) = 2 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Total segment failures (at master) = 0 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Total number of postmaster.pid files missing = 0 20181101:10:04:55:006659 hawq_state:test-1:gpadmin-[INFO]:-- Total number of postmaster.pid files found = 2 I have two questions about above log: 1) i have run "hawq init standby -n -M fast" but it still show "Standby is not synced" 2) standby node seems not work Does something I do wrong? Best Wishes From: "Lei Chang" <[email protected]> To: [email protected] Sent: Wednesday, October 31, 2018 8:34:16 PM Subject: Re: how to set standby after activate standby After activating standby as master, you can add a standby to the current master. Then you will have one master and one standby. Cheers Lei On Wed, Oct 31, 2018 at 5:51 PM < [email protected] > wrote: Hi! I test with hawq HA, and find if kill master process, activate standby as master is ok, but there is no more standby node: 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- HAWQ instance status summary 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:------------------------------------------------------ 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- Master instance = Active 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- No Standby master defined <==== 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- Total segment instance count from config file = 2 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:------------------------------------------------------ 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- Segment Status 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:------------------------------------------------------ 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- Total segments count from catalog = 2 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- Total segment valid (at master) = 2 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- Total segment failures (at master) = 0 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- Total number of postmaster.pid files missing = 0 20181031:17:44:05:1050788 hawq_state:test-1:gpadmin-[INFO]:-- Total number of postmaster.pid files found = 2 Does it mean HA only work for one time? Is there any way to set standby again after activate standby? Thanks
