View Issue Details

Jump to NotesJump to History
IDProjectCategoryView StatusDate SubmittedLast Update
0000517Pgpool-IIBugpublic2019-05-22 18:142019-05-23 16:54
Reporteradupa Assigned Topengbo  
PrioritynormalSeveritymajorReproducibilityalways
Status feedbackResolutionopen 
Product Version4.0.2 
Summary0000517: Healthcheck not starting for new node after reloading configuration
DescriptionI added a node to config file and run `pgpool reload`. New node is shown in `show pool_nodes;` but with status `down`.

Logs from reloading the config:

2019-05-22 08:51:03: pid 8: LOG: reload config files.
2019-05-22 08:51:03: pid 8: DEBUG: setting "backend_flag0" flag: 0002
2019-05-22 08:51:03: pid 8: DEBUG: initializing pool configuration
2019-05-22 08:51:03: pid 8: DETAIL: num_backends: 2 total_weight: 2.000000
2019-05-22 08:51:03: pid 8: DEBUG: initializing pool configuration
2019-05-22 08:51:03: pid 8: DETAIL: backend 0 weight: 1073741823.500000 flag: 0002
2019-05-22 08:51:03: pid 8: DEBUG: initializing pool configuration
2019-05-22 08:51:03: pid 8: DETAIL: backend 1 weight: 1073741823.500000 flag: 0000
2019-05-22 08:51:03: pid 60: LOG: reloading config file
2019-05-22 08:51:03: pid 60: DEBUG: setting "backend_flag0" flag: 0002
2019-05-22 08:51:03: pid 60: DEBUG: initializing pool configuration
2019-05-22 08:51:03: pid 60: DETAIL: num_backends: 2 total_weight: 2.000000
2019-05-22 08:51:03: pid 60: DEBUG: initializing pool configuration
2019-05-22 08:51:03: pid 60: DETAIL: backend 0 weight: 1073741823.500000 flag: 0002
2019-05-22 08:51:03: pid 60: DEBUG: initializing pool configuration
2019-05-22 08:51:03: pid 60: DETAIL: backend 1 weight: 1073741823.500000 flag: 0000

In ps -aux I see healthcheck process for backend 0, but not for backend 1

root 59 0.0 0.0 100212 3572 ? S 07:51 0:00 pgpool: PCP: wait for connection request
root 60 0.0 0.0 100208 4088 ? S 07:51 0:00 pgpool: worker process
root 61 0.0 0.0 100208 3660 ? S 07:51 0:00 pgpool: health check process(0)
root 62 0.0 0.0 115448 3440 pts/0 Ss 07:54 0:00 bash


When I start pgpool with the same configuration healthcheck process spawns correctly and both backends are up
TagsNo tags attached.

Activities

pengbo

2019-05-23 09:32

developer   ~0002624

Because Pgpool-II can't detect node automatically, you need to restart pgpool or attach the node to cluster by executing "pcp_attach_node".

Issue History

Date Modified Username Field Change
2019-05-22 18:14 adupa New Issue
2019-05-23 09:27 administrator Assigned To => pengbo
2019-05-23 09:27 administrator Status new => assigned
2019-05-23 09:32 pengbo Note Added: 0002624
2019-05-23 16:54 pengbo Status assigned => feedback