gpstart command fails to start the Greenplum Database

posted Apr 28, 2017, 4:42 PM by Sachchida Ojha
GPstart: gpstate failed. (Reason='could not connect to server: Connection refused)

To resolve the issue use the gpstop -af command on the master server and make sure there are no more postgres processes running.
Then do a gpstart.

[gpadmin@mdw]$ gpstart
[gpadmin@mdw]$ gpstate
The system is now started with all segments restored. 

If the gpstop -af command fails, you can kill the standby sync process on the smdw host.

[gpadmin@smdw]$ ps -ef |grep -i sync
root 4375 489 0 Nov22 ? 00:00:00 [xfssyncd]
gpadmin 15971 1 0 03:23 ? 00:00:00 /usr/local/greenplum-db/./bin/gpsyncmaster -D /data/master/gpseg-1 -p 5432 -b 98 -C -1 -z 48 -i
gpadmin 16147 16108 0 03:31 pts/0 00:00:00 grep -i sync

[gpadmin@smdw]$ kill -9 15971

[gpadmin@mdw]$ gpstart
[gpadmin@mdw]$ gpstop -r
[gpadmin@mdw]$ gprecoverseg
[gpadmin@mdw]$ gpstate