Skip to content

using pg_autoctl node_2 cannot bring back from maintenance mode #1074

@m1nx4s

Description

@m1nx4s

Hello,
I am using pg_auto_failover postgresql version 16.7, RHEL 8.10:
monitor node
node_1
node_2
When applying config changes in node_1, enabling maintenance in node_2:
pg_autoctl enable maintenance
When configuration completed in node_1, disabling maintenance:
pg_autoctl enable maintenance
This command worked fine for 5-7 times when did changes and node_2 catchedup
node_1 TLI: LSN.
When applied the last time enable, disable maintenance it somehow frozen and
received an error when tried to comeback from maintenance mode:
Name | Node | Host:Port | TLI: LSN | Connection |
Reported State | Assigned State
-------+-------+-------------------------+------------------+--------------+---------------------+--------------------
node_1 | 1 | node_1:5432 | 4: 97/6C000110 | read-write |
single | single
node_2 | 2 | node_2:5432 | 4: 97/6BD90750 | none ! |
maintenance | catchingup

The last known: TLI: LSN was 97/6BD90338
Name | Node | Host:Port | TLI: LSN | Connection |
Reported State | Assigned State
-------+-------+-------------------------+------------------+--------------+---------------------+--------------------
node_1 | 1 | node_1:5432 | 4: 97/6BD90338 | read-write |
primary | primary
node_2 | 2 | node_2:5432 | 4: 97/6BD90338 | read-only |
secondary | secondary

This happened for me twice in 2 days after several commands when
enabled/disabled maintenance mode.

Is there any known issue why this happens randomly for node_2.
The only wait how i was able to fix it, i had to drop node_2 from monitor
node and reapply setup on node_2.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions