[pgpool-general: 2918] Re: pgpool-II V3.3.3 will not start with use_watchdog = on
nagata at sraoss.co.jp
Sun Jun 15 01:03:46 JST 2014
On Fri, 13 Jun 2014 09:19:36 -0400
John Scalia <jayknowsunix at gmail.com> wrote:
> Konichiwa Nagata-san,
> I have finally succeeded in getting all pgpool's started on the 3 machines. In order to send you their individual configurations, I first needed to collect these on our publicly
> accessible jumphost server. There I renamed them to pgpool.conf.primary and pgpool.conf.standby1, etc. Then, I performed a diff between them, first comparing the primary against
> standby1. One thing looked suspicious and I discovered a minor typo in the primary's configuration. Then I compared the primary against standby2. Many more items were suspicious. I
> copied the primary's pgpool.conf file to standby2, and made the necessary modifications in the watchdog area. After doing this, everything started correctly on all 3 systems.
> I do, however, have a question, immediately following the wd_lifecheck area is a section for "Other pgpool Connection Settings", and I must confess that I really don't understand
> how to set these. Allow me to explain that my servers are set as:
> primary: 10.10.1.128
> standby1: 10.10.1.129
> standby2: 10.10.1.17
> delegate_IP: 10.10.1.130
> Now, on the primary, I currently have the pgpool.conf setting for other_pgpool_hostname0 = 10.10.1.129 and other_pgpool_hostname1 = 10.10.1.17 while on standby1 these settings are
> other_pgpool_hostname0 = 10.10.1.128 and other_pgpool_hostname1 = 10.10.1.17. On the second standby server, other_pgpool_hostname0 = 10.10.1.128 and other_pgpool_hostname1 =
> This configuration is starting up acceptably, but I really don't know that I have this portion correct. Can you advise?
No problem, The configuration is correct. You can spcify
other_pgpool_hostnameX in any order.
> On 6/12/2014 10:19 PM, Yugo Nagata wrote:
> > Hi,
> > pgpool for standby will fails to start when the configuration of virtual IP
> > (delegate_IP) is different from active pgpool. Could you please see if
> > there is any error messages in active pgpool.
> > If possible, could you please send me log messages and pgpool.confs of
> > all pgpools. These would be helpful to resolve.
> > On Thu, 12 Jun 2014 08:45:32 -0400
> > John Scalia <jayknowsunix at gmail.com> wrote:
> >> Greetings all,
> >> I'm fairly new to pgpool-II, and I'm still trying to learn about all it's options, but my three test servers, configured as a primary with 2 standby servers in a master-slave
> >> arrangement, have developed a problem. Pgpool-II starts and runs just fine on the primary database server, but it will not start on either standby if I have use_watchdog = on. The
> >> log reports:
> >> LOG: pid 23024: wd_check_setuid all commands have setuid bit
> >> LOG: pid 23024: watchdog might call network commands which using setuid bit.
> >> ERROR: pid 23024: wd_init: failed to start watchdog
> >> ERROR: pid 23024: watchdog: wd_init_failed
> >> ERROR: pid 23024: wd_main error
> >> ERROR: pid 23024: unlink(/tmp/.s.PGSQL.9898) failed: No such file or directory
> >> I'm fairly sure the last line from the log is just a cleanup attempt which fails as the socket never got created, but otherwise, I'm not sure why this is actually failing. On rare
> >> occasion pgpool-II will sort of start, but none of the child processes get spawned, and a ps -ef | grep pgpool yields only the:
> >> /usr/bin/pgpool -f /etc/pgpool-II/pgpool.conf -n
> >> line. Now, if I can't get this resolved, what exactly am I losing by not running watchdog? Of course, I'd really like to have this working in its entirety.
> >> Thanks in advance,
> >> Jay
> >> _______________________________________________
> >> pgpool-general mailing list
> >> pgpool-general at pgpool.net
> >> http://www.pgpool.net/mailman/listinfo/pgpool-general
Yugo Nagata <nagata at sraoss.co.jp>
More information about the pgpool-general