[postgres-xl-bugs] Coordinator restart is not the same as stop and start

Julian Szafraniec findingfreeusernamesishard at gmail.com
Mon Aug 21 08:44:05 PDT 2017


Hi guys,

I think there is something wrong with `pg_ctl restart` command.
When setting up the cluster using pgxc_ctl, all servers will bind to
0.0.0.0 address instead of the one specified in the config file.

Therefore I alter postgres.conf file manually to bind it to specific interface.
When I issue `pg_ctl restart -w -Z coordinator -D data_dir` it will
not bind to the specified address.
On the other hand if I issue a separate `pg_ctl stop` and `pg_ctl
start` it will bind fine.
I think it is either a bug or a really weird feature:)

Additionally there is something wrong with the command itself.
It will not finish cleanly even when run over ssh.
For example executing:
```
ssh 192.168.11.1 -C "pg_ctl restart -w -Z coordinator -D
/var/lib/postgres/data_c01"
```
will hang with this as the result:
```
LOG:  redirecting log output to logging collector process
HINT:  Future log output will appear in directory "pg_log".
```
Regards
Julian


More information about the postgres-xl-bugs mailing list