View Issue Details

IDProjectCategoryView StatusLast Update
0000514Pgpool-IIBugpublic2019-05-23 16:30
ReportervanAssigned Topengbo 
PrioritynormalSeveritymajorReproducibilityalways
Status resolvedResolutionopen 
Product Version4.0.2 
Target VersionFixed in Version 
Summary0000514: verify_backend_node_status: there's no primary node
DescriptionHi , i tested with just one node (Primary )

After restarting the primary node, pgpool is not able to pick the node.



2019-05-20 09:25:19: pid 12: DEBUG: verify_backend_node_status: there's no primary node
2019-05-20 09:25:20: pid 214: DEBUG: pool_flush_it: flush size: 40
2019-05-20 09:25:20: pid 214: DEBUG: pool_read: read 326 bytes from backend 0
2019-05-20 09:25:20: pid 214: DEBUG: authenticate kind = 0
2019-05-20 09:25:20: pid 214: DEBUG: authenticate backend: key data received
2019-05-20 09:25:20: pid 214: DEBUG: authenticate backend: transaction state: I
2019-05-20 09:25:20: pid 214: DEBUG: do_query: extended:0 query:"SELECT pg_is_in_recovery()"
2019-05-20 09:25:20: pid 214: DEBUG: pool_write: to backend: 0 kind:Q
2019-05-20 09:25:20: pid 214: DEBUG: pool_flush_it: flush size: 32
2019-05-20 09:25:20: pid 214: DEBUG: pool_read: read 75 bytes from backend 0
2019-05-20 09:25:20: pid 214: DEBUG: do_query: kind: 'T'
2019-05-20 09:25:20: pid 214: DEBUG: do_query: received ROW DESCRIPTION ('T')
2019-05-20 09:25:20: pid 214: DEBUG: do_query: row description: num_fileds: 1
2019-05-20 09:25:20: pid 214: DEBUG: do_query: kind: 'D'
2019-05-20 09:25:20: pid 214: DEBUG: do_query: received DATA ROW ('D')
2019-05-20 09:25:20: pid 214: DEBUG: do_query: kind: 'C'


TagsNo tags attached.

Activities

van

2019-05-20 18:33

reporter  

pgpool_logs (393,312 bytes)

pengbo

2019-05-21 14:19

developer   ~0002613

Last edited: 2019-05-21 16:33

View 2 revisions

When you stop a node, the node will be detached from cluster.
Even if you start the node, the node will never return to the previous state.

So, after you restart a node, you need to attach this node to cluster or restart pgpool by using "-D" option.

(1) pcp_attach_node -p xxxx -U xxxx -n 0
See docs: http://www.pgpool.net/docs/latest/ja/html/pcp-attach-node.html

OR

(2) If you install pgpool by using RPM, add "-D" option to /etc/sysconfig/pgpool
---
OPTS=" -D -n"
---

van

2019-05-21 16:04

reporter   ~0002615

Hi pengbo,

Sorry i am new to pgpool, when you say restart , you mean stop the pgpool and start it again ?
or some just reconfiguration with some command?

pengbo

2019-05-21 16:35

developer   ~0002619

Last edited: 2019-05-21 16:35

View 2 revisions

If you installed pgpool by using RPM, try the 2 steps below:

(1) Add "-D" option to "/etc/sysconfig/pgpool"
---
OPTS=" -D -n"
---
(2) Restart pgpool by using the command below:

systemctl restart pgpool

van

2019-05-21 18:27

reporter   ~0002621

Hi pengbo,

I install it from tar.

My question is that

1) I ran the pgpool with -D option
2) make DB node down
3) pgpool DB node status down
4)make DB node UP
5) pgpool DB node still status down

does pgpool should automatically detect that db node UP and make status UP ?
or We have to run pcp_attach_node command manually to make it UP?

pengbo

2019-05-21 19:14

developer   ~0002622

| does pgpool should automatically detect that db node UP and make status UP ?
| or We have to run pcp_attach_node command manually to make it UP?

Pgpool-II can't automaticlly detect the DB node that is DOWN.
So, you have to run pcp_attach_node manually.

van

2019-05-23 13:00

reporter   ~0002626

Thanks pengbo .

pengbo

2019-05-23 13:13

developer   ~0002627

If you have resolved this issue, could you mark this issue as "resolved"?

pengbo

2019-05-23 13:19

developer   ~0002628

BTW, in next major release version Pgpool-II 4.1, a new featue "auto failback " will be added.
With this feature, Pgpool-II will automatically bring back a "healthy" standby node.

See more details:
https://www.pgpool.net/pipermail/pgpool-hackers/2019-May/003328.html

van

2019-05-23 14:29

reporter   ~0002631

Hi pengbo,

"auto failback " -> that will be really helpful :) .

My issue is resolved now but i have no idea "how to mark it as resolved".

pengbo

2019-05-23 16:30

developer   ~0002632

Mark this issue as "resolved" status.

Issue History

Date Modified Username Field Change
2019-05-20 18:32 van New Issue
2019-05-20 18:33 van File Added: pgpool_logs
2019-05-21 11:01 administrator Assigned To => pengbo
2019-05-21 11:01 administrator Status new => assigned
2019-05-21 14:19 pengbo Note Added: 0002613
2019-05-21 14:19 pengbo Status assigned => feedback
2019-05-21 16:04 van Note Added: 0002615
2019-05-21 16:04 van Status feedback => assigned
2019-05-21 16:33 administrator Note Edited: 0002613 View Revisions
2019-05-21 16:35 pengbo Note Added: 0002619
2019-05-21 16:35 pengbo Note Edited: 0002619 View Revisions
2019-05-21 18:27 van Note Added: 0002621
2019-05-21 19:14 pengbo Note Added: 0002622
2019-05-23 13:00 van Note Added: 0002626
2019-05-23 13:13 pengbo Note Added: 0002627
2019-05-23 13:19 pengbo Note Added: 0002628
2019-05-23 14:29 van Note Added: 0002631
2019-05-23 16:30 pengbo Status assigned => resolved
2019-05-23 16:30 pengbo Note Added: 0002632