[pgpool-general: 4900] Re: pgpool duplicate IP (3.5.3)

Fri Aug 12 03:16:30 JST 2016

Hi,
still after modifing the setuid bit on 'ifconfig' and 'arping' commands,
when rebooting the primary node the 2nd node gets promoted (vip) and the
faulty node boots up as primary:

tail: /var/log/pgpool.log: file truncated
2016-08-11 20:45:40: pid 1789: LOG:  reading status file: 0 th backend is
set to down status
2016-08-11 20:45:40: pid 1789: LOG:  waiting for watchdog to initialize
2016-08-11 20:45:40: pid 1795: LOG:  setting the local watchdog node name
to "Linux_mgrdb84_9999"
2016-08-11 20:45:40: pid 1795: LOG:  watchdog cluster configured with 1
remote nodes
2016-08-11 20:45:40: pid 1795: LOG:  watchdog remote node:0 on 1.1.1.85:9000
2016-08-11 20:45:40: pid 1795: LOG:  interface monitoring is disabled in
watchdog
*2016-08-11 20:45:40: pid 1795: LOG:  IPC socket path:
"/tmp/.s.PGPOOLWD_CMD.9000"*
*2016-08-11 20:45:45: pid 1795: LOG:  watchdog node state changed from
[LOADING] to [JOINING]*
*2016-08-11 20:45:50: pid 1795: LOG:  watchdog node state changed from
[JOINING] to [INITIALIZING]*
*2016-08-11 20:45:51: pid 1795: LOG:  I am the only alive node in the
watchdog cluster*
*2016-08-11 20:45:51: pid 1795: HINT:  skiping stand for coordinator state*
*2016-08-11 20:45:51: pid 1795: LOG:  watchdog node state changed from
[INITIALIZING] to [MASTER]*
*2016-08-11 20:45:51: pid 1795: LOG:  I am announcing my self as
master/coordinator watchdog node*
*2016-08-11 20:45:54: pid 1795: LOG:  new outbond connection to
1.1.1.85:9000 <http://1.1.1.85:9000>*
2016-08-11 20:45:56: pid 1795: LOG:  I am the cluster leader node
2016-08-11 20:45:56: pid 1795: DETAIL:  our declare coordinator message is
accepted by all nodes
2016-08-11 20:45:56: pid 1795: LOG:  I am the cluster leader node. Starting
escalation process
2016-08-11 20:45:56: pid 1789: LOG:  watchdog process is initialized
2016-08-11 20:45:56: pid 1795: LOG:  escalation process started with
PID:2130
2016-08-11 20:45:56: pid 2130: LOG:  watchdog: escalation started
2016-08-11 20:45:56: pid 1789: LOG:  Setting up socket for 0.0.0.0:9999
2016-08-11 20:45:56: pid 1789: LOG:  Setting up socket for :::9999
2016-08-11 20:45:56: pid 2131: LOG:  2 watchdog nodes are configured for
lifecheck
2016-08-11 20:45:56: pid 2131: LOG:  watchdog nodes ID:0
Name:"Linux_mgrdb84_9999"
2016-08-11 20:45:56: pid 2131: DETAIL:  Host:"1.1.1.84" WD Port:9000
pgpool-II port:9999
2016-08-11 20:45:56: pid 2131: LOG:  watchdog nodes ID:1 Name:"Not_Set"
2016-08-11 20:45:56: pid 2131: DETAIL:  Host:"1.1.1.85" WD Port:9000
pgpool-II port:9999
2016-08-11 20:45:56: pid 1789: LOG:  pgpool-II successfully started.
version 3.5.3 (ekieboshi)
2016-08-11 20:45:56: pid 1789: LOG:  find_primary_node: checking backend no
0
2016-08-11 20:45:56: pid 1789: LOG:  find_primary_node: checking backend no
1
2016-08-11 20:45:56: pid 1789: LOG:  find_primary_node: primary node id is 1
2016-08-11 20:45:57: pid 2135: LOG:  createing watchdog heartbeat receive
socket.
2016-08-11 20:45:57: pid 2135: DETAIL:  bind receive socket to device:
"eth1"
2016-08-11 20:45:57: pid 2135: LOG:  set SO_REUSEPORT option to the socket
2016-08-11 20:45:57: pid 2135: LOG:  creating watchdog heartbeat receive
socket.
2016-08-11 20:45:57: pid 2135: DETAIL:  set SO_REUSEPORT
2016-08-11 20:45:57: pid 2137: LOG:  creating socket for sending heartbeat
2016-08-11 20:45:57: pid 2137: DETAIL:  bind send socket to device: eth1
2016-08-11 20:45:57: pid 2137: LOG:  set SO_REUSEPORT option to the socket
2016-08-11 20:45:57: pid 2137: LOG:  creating socket for sending heartbeat
2016-08-11 20:45:57: pid 2137: DETAIL:  set SO_REUSEPORT
2016-08-11 20:45:58: pid 2130: WARNING:  watchdog failed to bring up
delegate IP, 'if_up_cmd' failed
2016-08-11 20:45:58: pid 2130: WARNING:  watchdog de-escalation failed to
bring down delegate IP
2016-08-11 20:45:58: pid 1795: LOG:  watchdog escalation process with pid:
2130 exit with SUCCESS.
2016-08-11 20:47:24: pid 1795: LOG:  new watchdog node connection is
received from "1.1.1.85:17053"
2016-08-11 20:47:36: pid 2131: LOG:  watchdog: lifecheck started

please advise...
cohavisi

On Thu, Aug 11, 2016 at 6:31 PM, Daniel Huhardeaux <tech at tootai.net> wrote:

> Hello
>
> Le 11/08/2016 à 16:46, Shay Cohavi a écrit :
>
>> Hi,
>>
>> When performing an restart on a primary pgpool node, and vip transfer to
>> the 2nd node, but when the faulty primary boots up, it  declare itself
>> as the only node in the cluster and brings up the VIP (duplicate IP)!!
>>
>>
>> the 1st node (startup):
>> 2016-08-11 17:31:36: pid 1761: WARNING:  checking setuid bit of if_up_cmd
>> 2016-08-11 17:31:36: pid 1761: DETAIL:  ifup[/sbin/ifconfig] doesn't
>> have setuid bit
>> 2016-08-11 17:31:36: pid 1761: WARNING:  checking setuid bit of
>> if_down_cmd
>> 2016-08-11 17:31:36: pid 1761: DETAIL:  ifdown[/sbin/ifconfig] doesn't
>> have setuid bit
>> 2016-08-11 17:31:36: pid 1761: WARNING:  checking setuid bit of arping
>> command
>> 2016-08-11 17:31:36: pid 1761: DETAIL:  arping[/sbin/arping] doesn't
>> have setuid bit
>>
>
> Answer to your problem is here, set setuid bit
>
>
>> the 2nd node:
>>
>> *2016-08-11 17:17:34: pid 16256: WARNING:  checking setuid bit of
>> if_up_cmd*
>> 2016-08-11 17:17:34: pid 16256: DETAIL:  ifup[/sbin/ifconfig] doesn't
>> have setuid bit
>> *2016-08-11 17:17:34: pid 16256: WARNING:  checking setuid bit of
>> if_down_cmd*
>> 2016-08-11 17:17:34: pid 16256: DETAIL:  ifdown[/sbin/ifconfig] doesn't
>> have setuid bit
>> *2016-08-11 17:17:34: pid 16256: WARNING:  checking setuid bit of arping
>> command*
>> 2016-08-11 17:17:34: pid 16256: DETAIL:  arping[/sbin/arping] doesn't
>> have setuid bit
>>
>
> Same here
>
> BTW i hope that both servers are not connected to Internet, IP range you
> are using are not in private range, they belong to APNIC.
>
> Regards
>
> Daniel
> --
> TOOTAi Networks
> _______________________________________________
> pgpool-general mailing list
> pgpool-general at pgpool.net
> http://www.pgpool.net/mailman/listinfo/pgpool-general
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20160811/eb0d9d57/attachment-0001.html>