[pgpool-general: 1348] Re: [PgpoolI 3.2.1] Watchdog: delegate_IP not bind by pgpool.

[Thomas Martin]

Hi Wolf.

Thanks for your answer.

Yes I'm using pgpool with a local postgres as backend (and my second
pgpool node have the second postgresql node).

Also you are right: if I use the wildcard "*" then pgpool will bind on
everything on the server. But in that case I will have to put postgres
on a other port (I know in my example postgres is listening on 15432
but in a production environment I should have postgres binding 5432).
Do you know if the utilization of the wilcard is absolutely mandatory ?

Regards.

2013/1/28 Wolf Schwurack <wolf at uen.org>:
> My setup I have 2 pgpool servers and master/slave postgres servers.  Are you running pgpool on the same server as postgres? You may want to try opening to all with the listen_addresses first and leave wd_hostname blank to see if that will work
>
> Here's the info on my pgpool.conf
>
> listen_addresses = '*'  --  open to all
>  port = 9999
>
>  pcp_port = 9898
>
>  backend_hostname0 = 'pgtest'    -- postgres server
>  backend_port0 = 5432
>  backend_weight0 = 1
>  backend_data_directory0 = '/data'
>  backend_flag0 = 'ALLOW_TO_FAILOVER'
>  backend_hostname1 = 'pgtest2'  -- postgres server
>  backend_port1 = 5432
>  backend_weight1 = 1
>  backend_data_directory1 = '/data'
>  backend_flag1 = 'ALLOW_TO_FAILOVER'
>
>  connection_cache = on
>  load_balance_mode = on
>
>  use_watchdog = on
>  trusted_servers = ''
>  delegate_IP = '10.10.0.20'  -- watchdog ip
>  wd_hostname = ' '
>  wd_port = 9000
>  wd_interval = 10
>
> Wolf
>
> -----Original Message-----
> From: pgpool-general-bounces at pgpool.net [mailto:pgpool-general-bounces at pgpool.net] On Behalf Of Thomas Martin
> Sent: Monday, January 28, 2013 6:16 AM
> To: pgpool-general at pgpool.net
> Subject: [pgpool-general: 1346] [PgpoolI 3.2.1] Watchdog: delegate_IP not bind by pgpool.
>
> Hello everyone.
>
> I guess I'm having an issue with watchdog.
> Here is the thing: the delegate_IP configured is not bind by pgpool (but mount correctly by watchdog).
>
> I read the documentation
> (http://www.pgpool.net/docs/pgpool-II-3.2.1/pgpool-en.html#watchdog)
> and the tutorial
> (http://www.pgpool.net/pgpool-web/contrib_docs/watchdog/en.html) but I can't figure out what I'm missing.
>
> I have two nodes: 10.59.10.65 and 10.59.10.66 (the delegate IP is 10.59.10.67).
> Here is my configuration on the first node (10.59.10.65):
>  listen_addresses = '10.59.10.65'
>  port = 5432
>
>  pcp_port = 9898
>
>  backend_hostname0 = '10.59.10.65'
>  backend_port0 = 15432
>  backend_weight0 = 1
>  backend_data_directory0 = '/data'
>  backend_flag0 = 'ALLOW_TO_FAILOVER'
>  backend_hostname1 = '10.59.10.66'
>  backend_port1 = 15432
>  backend_weight1 = 1
>  backend_data_directory1 = '/data'
>  backend_flag1 = 'ALLOW_TO_FAILOVER'
>
>  connection_cache = on
>  load_balance_mode = on
>
>  use_watchdog = on
>  trusted_servers = ''
>  delegate_IP = '10.59.10.67'
>  wd_hostname = '10.59.10.65'
>  wd_port = 9000
>  wd_interval = 10
>
> And here is my logs when I start the pgpool:
> root at pgpool2-1:/etc/itf-pgpool-ii# /usr/bin/itf/postgres/pgpool -n
> 2013-01-28 13:03:15 LOG:   pid 3608: wd_chk_sticky:
> ifup[/sbin/ifconfig] doesn't have sticky bit
> 2013-01-28 13:03:15 LOG:   pid 3608: wd_create_send_socket: connect()
> reports failure (Connection refused). You can safely ignore this while starting up.
> 2013-01-28 13:03:18 LOG:   pid 3608: wd_escalation: escalated to master pgpool
> 2013-01-28 13:03:18 LOG:   pid 3608: wd_create_send_socket: connect()
> reports failure (Connection refused). You can safely ignore this while starting up.
> 2013-01-28 13:03:18 LOG:   pid 3608: wd_escalation:  escaleted to
> delegate_IP holder
> 2013-01-28 13:03:18 LOG:   pid 3608: wd_init: start watchdog
> 2013-01-28 13:03:18 LOG:   pid 3608: pgpool-II successfully started.
> version 3.2.1 (namameboshi)
> 2013-01-28 13:03:19 LOG:   pid 3625: connection received:
> host=10.59.10.65 port=42060
>
>
> The ip is correctly mounted:
> root at pgpool2-1:~# ifconfig eth0:0
> eth0:0    Link encap:Ethernet  HWaddr 00:16:3e:27:bd:89
>           inet addr:10.59.10.67  Bcast:10.59.10.255  Mask:255.255.255.0
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           Interrupt:245
>
>
> But nothing bind it:
> root at pgpool2-1:~# netstat -lnp|grep -E "pgpool|watchdog|10.59.10.67"
> tcp        0      0 10.59.10.65:5432        0.0.0.0:*
> LISTEN      3608/pgpool
> tcp        0      0 0.0.0.0:9000            0.0.0.0:*
> LISTEN      3613/pgpool: watchd
> tcp        0      0 0.0.0.0:9898            0.0.0.0:*
> LISTEN      3608/pgpool
> unix  2      [ ACC ]     STREAM     LISTENING     20995    3608/pgpool
>         /tmp/.s.PGSQL.5432
> unix  2      [ ACC ]     STREAM     LISTENING     21023    3608/pgpool
>         /tmp/.s.PGSQL.9898
>
>
> I guess I just misunderstood how watchdog works !?
> How am I supposed to have the pgpool binding 10.59.10.67 ? Does I have to make an iptables rules ?
>
>
> Thanks.
> _______________________________________________
> pgpool-general mailing list
> pgpool-general at pgpool.net
> http://www.pgpool.net/mailman/listinfo/pgpool-general