[pgpool-general: 7019] Re: Pgpool-II-4.1.1 and Ubuntu 20.04
Wolf Schwurack
wolf at uen.org
Fri May 15 05:39:03 JST 2020
Again more testing - When these are commented out
#other_pgpool_hostname1 = 'pgtest-03'
#other_pgpool_port1 = 9999
#other_wd_port1 = 9000
The third node fails to start pgpool
On 5/14/20, 2:10 PM, "pgpool-general-bounces at pgpool.net on behalf of Wolf Schwurack" <pgpool-general-bounces at pgpool.net on behalf of wolf at uen.org> wrote:
After some more testing I found that if I comment out "#other_pgpool_hostname1" parameters in pgpool.conf, both version 4.0.8 and 4.1.1 watchdog starts with just the master node started
# - Other pgpool Connection Settings -
other_pgpool_hostname0 = 'pgtest-02'
other_pgpool_port0 = 9999
other_wd_port0 = 9000
#other_pgpool_hostname1 = 'pgtest-03'
#other_pgpool_port1 = 9999
#other_wd_port1 = 9000
pgpool.log
2020-05-14 13:57:20: pid 52482: LOG: successfully acquired the delegate IP:"10.11.0.204"
2020-05-14 13:57:20: pid 52482: DETAIL: 'if_up_cmd' returned with success
2020-05-14 13:57:20: pid 52476: LOG: watchdog escalation process with pid: 52482 exit with SUCCESS.
When I enable these parameters in pgpool.conf
other_pgpool_hostname1 = 'pgtest-03'
other_pgpool_port1 = 9999
other_wd_port1 = 9000
The master node fails to start watchdog until another pgpool node is started.
Wolfgang Schwurack
Database/System Administrator
Utah Education Network
801-587-9444
wolf at uen.org
On 5/13/20, 10:34 AM, "pgpool-general-bounces at pgpool.net on behalf of Wolf Schwurack" <pgpool-general-bounces at pgpool.net on behalf of wolf at uen.org> wrote:
I installed pgpool-II-4.0.8 using pgpool.conf from pgpool-II-4.1.1 watchdog failed to start. But when I used pgpool.conf from pgpool-II-4.0.8 watchdog started on a single node
2020-05-13 10:28:50: pid 11671: LOG: pgpool-II successfully started. version 4.0.8 (torokiboshi)
2020-05-13 10:28:50: pid 11671: LOG: node status[0]: 1
2020-05-13 10:28:50: pid 11671: LOG: node status[1]: 2
2020-05-13 10:28:51: pid 11685: LOG: creating socket for sending heartbeat
2020-05-13 10:28:51: pid 11685: DETAIL: bind send socket to device: ens160
2020-05-13 10:28:51: pid 11685: LOG: set SO_REUSEPORT option to the socket
2020-05-13 10:28:51: pid 11685: LOG: creating socket for sending heartbeat
2020-05-13 10:28:51: pid 11685: DETAIL: set SO_REUSEPORT
2020-05-13 10:28:51: pid 11684: LOG: createing watchdog heartbeat receive socket.
2020-05-13 10:28:51: pid 11684: DETAIL: bind receive socket to device: "ens160"
2020-05-13 10:28:51: pid 11684: LOG: set SO_REUSEPORT option to the socket
2020-05-13 10:28:51: pid 11684: LOG: creating watchdog heartbeat receive socket.
2020-05-13 10:28:51: pid 11684: DETAIL: set SO_REUSEPORT
2020-05-13 10:28:57: pid 11681: LOG: successfully acquired the delegate IP:"10.11.0.204"
2020-05-13 10:28:57: pid 11681: DETAIL: 'if_up_cmd' returned with success
2020-05-13 10:28:57: pid 11673: LOG: watchdog escalation process with pid: 11681 exit with SUCCESS.
Wolfgang Schwurack
Database/System Administrator
Utah Education Network
801-587-9444
wolf at uen.org
On 5/13/20, 9:39 AM, "pgpool-general-bounces at pgpool.net on behalf of Wolf Schwurack" <pgpool-general-bounces at pgpool.net on behalf of wolf at uen.org> wrote:
Still testing this I decided to start pgpool on the second node which started watchdog. I know you told me before about having an even number a nodes to set "enable_consensus_with_half_votes = on" Which is how this is set in pgpool.conf. But I have 3 nodes setup which is uneven number of nodes.
It seems that the parameter "enable_consensus_with_half_votes" is not working. I can set it to 'on or off' The delegate IP is only successfully acquired if the second node is running pgpool
Here's the log output once I start pgpool on the second node.
2020-05-13 09:11:08: pid 4103: LOG: new watchdog node connection is received from "10.11.0.206:17581"
2020-05-13 09:11:08: pid 4103: LOG: new node joined the cluster hostname:"pgtest-02" port:9000 pgpool_port:9999
2020-05-13 09:11:08: pid 4103: DETAIL: Pgpool-II version:"4.1.1" watchdog messaging version: 1.1
2020-05-13 09:11:08: pid 4103: LOG: new outbound connection to pgtest-02:9000
2020-05-13 09:11:14: pid 4103: LOG: adding watchdog node "pgtest-02:9999 Linux pgtest-02" to the standby list
2020-05-13 09:11:14: pid 4103: LOG: quorum found
2020-05-13 09:11:14: pid 4103: DETAIL: starting escalation process
2020-05-13 09:11:14: pid 4103: LOG: escalation process started with PID:4231
2020-05-13 09:11:14: pid 4099: LOG: Pgpool-II parent process received watchdog quorum change signal from watchdog
2020-05-13 09:11:14: pid 4103: LOG: new IPC connection received
2020-05-13 09:11:14: pid 4099: LOG: watchdog cluster now holds the quorum
2020-05-13 09:11:14: pid 4099: DETAIL: updating the state of quarantine backend nodes
2020-05-13 09:11:14: pid 4103: LOG: new IPC connection received
2020-05-13 09:11:14: pid 4231: LOG: watchdog: escalation started
2020-05-13 09:11:21: pid 4231: LOG: successfully acquired the delegate IP:"10.11.0.204"
2020-05-13 09:11:21: pid 4231: DETAIL: 'if_up_cmd' returned with success
2020-05-13 09:11:21: pid 4103: LOG: watchdog escalation process with pid: 4231 exit with SUCCESS.
Then I start pgpool on the third node
2020-05-13 09:26:52: pid 4906: LOG: new watchdog node connection is received from "10.11.0.207:47278"
2020-05-13 09:26:52: pid 4906: LOG: new node joined the cluster hostname:"pgtest-03" port:9000 pgpool_port:9999
2020-05-13 09:26:52: pid 4906: DETAIL: Pgpool-II version:"4.1.1" watchdog messaging version: 1.1
2020-05-13 09:26:52: pid 4906: LOG: new outbound connection to pgtest-03:9000
2020-05-13 09:26:58: pid 4906: LOG: adding watchdog node "pgtest-03:9999 Linux pgtest-03" to the standby list
Now when I stop pgpool on third node
2020-05-13 09:34:19: pid 4906: LOG: client socket of pgtest-03:9999 Linux pgtest-03 is closed
2020-05-13 09:34:19: pid 4906: LOG: remote node "pgtest-03:9999 Linux pgtest-03" is shutting down
2020-05-13 09:34:19: pid 4906: LOG: removing watchdog node "pgtest-03:9999 Linux pgtest-03" from the standby list
Watchdog is still running
root at pgtest-01:~# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether 00:50:56:b3:d3:58 brd ff:ff:ff:ff:ff:ff
inet 10.11.0.205/24 brd 10.11.0.255 scope global ens160
valid_lft forever preferred_lft forever
inet 10.11.0.204/24 brd 10.11.0.255 scope global secondary ens160:1
valid_lft forever preferred_lft forever
Now when I stop pgpool on the second node
2020-05-13 09:36:01: pid 4906: LOG: We have lost the quorum
2020-05-13 09:36:01: pid 4902: LOG: Pgpool-II parent process received watchdog quorum change signal from watchdog
2020-05-13 09:36:01: pid 4906: LOG: new IPC connection received
2020-05-13 09:36:01: pid 5504: LOG: watchdog: de-escalation started
2020-05-13 09:36:01: pid 5504: LOG: successfully released the delegate IP:"10.11.0.204"
2020-05-13 09:36:01: pid 5504: DETAIL: 'if_down_cmd' returned with success
2020-05-13 09:36:01: pid 4906: LOG: watchdog de-escalation process with pid: 5504 exit with SUCCESS.
_______________________________________________
pgpool-general mailing list
pgpool-general at pgpool.net
http://www.pgpool.net/mailman/listinfo/pgpool-general
_______________________________________________
pgpool-general mailing list
pgpool-general at pgpool.net
http://www.pgpool.net/mailman/listinfo/pgpool-general
_______________________________________________
pgpool-general mailing list
pgpool-general at pgpool.net
http://www.pgpool.net/mailman/listinfo/pgpool-general
More information about the pgpool-general
mailing list