View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0000586||Pgpool-II||[All Projects] General||public||2020-02-26 17:47||2020-04-14 13:23|
|Priority||normal||Severity||minor||Reproducibility||have not tried|
|Target Version||Fixed in Version|
|Summary||0000586: Do not update status of slave host on PGPool|
|Description||Dear pgPool team,|
We are using standard master slave scenario with pgpool 4.1.1.
Currently, we have a problem as bellow:
① master (on) + slave (on) : status on pgpool => up -up
② master (on) + slave (off) : status on pgpool => up-down
③ master(on) + slave(re-on): status on pgpool => up-down
Could you please tell me about that do having an setting on pgpool.conf to update status of node on pgpool ?
or that is abnormal point on latest version 4.1.1 of PgPool ?
Thank you very much for your support !
|Tags||No tags attached.|
> master(on) + slave(re-on): status on pgpool => up-down
> Could you please tell me about that do having an setting on pgpool.conf to update status of node on pgpool ?
> or that is abnormal point on latest version 4.1.1 of PgPool ?
If you mean by "re-on" starting the slave again, then the behavior you are seeing is normal since Pgpool-II was born and explicitly stated in the doc:
"A PostgreSQL node detached by failover or switch over will never return to the previous state (attached state). "
In this you can use pcp_attach_node to make the slave to be in "up" as explained in the doc.
Or you can configure "auto_failback" which is new in 4.1 without issuing pcp_attach_node command. See the doc for more details.
Dear Mr t-isshi,
Thank you very much for your support,
I has been using pcp_attach_node as bellow
> pcp_recovery_node 60 10.116.226.186 9898 postgres postgres 1
But, this is current status of node server when using command "show pool_nodes" to show
0 | 10.116.226.63 | 5432 | up | 0.500000 | primary | 4
| true | 0 | |
| 2020-02-26 15:29:14
1 | 10.116.226.155 | 5432 | quarantine | 0.500000 | standby | 0
| false | 0 | streaming | async
| 2020-02-26 16:42:30
=> Status of slave is change from "down" to "quarantine ".
On the other way, I using "pcp_node_info" to show status of 'slave' node. But having an other result - 'status is down 'as :
-bash-4.2$ pcp_node_info -h localhost -U postgres 1
10.116.226.155 5432 3 0.500000 down standby 0 streaming async
Having any abnormal point on this case ?
I'm sorry but pcp_attach_node command that i using as : pcp_attach_node
pcp_attach_node -U postgres -h 10.51.219.14 -p 9898 1
You need to turn on enable_consensus_with_half_votes because you have only 2 Pgpool-II nodes. We recommend to have at least 3 Pgpool-II nodes.
See release note (https://www.pgpool.net/docs/latest/en/html/release-4-1-0.html) for more details.
Thank you very much for support.
I stop pgpool services and delete file /tmp/pgpool_status.
After that restart pgool services and re-run this command:
> pcp_attach_node -U postgres -h 10.51.219.14 -p 9898 1
And it work ok with my case.
||May I close this issue?|
|2020-02-26 17:47||ducpham||New Issue|
|2020-02-27 09:02||t-ishii||Note Added: 0003226|
|2020-02-27 10:57||ducpham||Note Added: 0003228|
|2020-02-27 12:13||ducpham||Note Added: 0003229|
|2020-02-27 13:29||t-ishii||Note Added: 0003230|
|2020-03-02 11:22||t-ishii||Status||new => feedback|
|2020-03-02 11:22||t-ishii||Assigned To||=> t-ishii|
|2020-03-02 11:22||t-ishii||Status||feedback => assigned|
|2020-03-03 13:36||t-ishii||Status||assigned => feedback|
|2020-03-03 15:47||ducpham||Note Added: 0003236|
|2020-03-03 15:47||ducpham||Status||feedback => assigned|
|2020-04-14 13:23||t-ishii||Note Added: 0003318|
|2020-04-14 13:23||t-ishii||Status||assigned => feedback|