View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000255 | Pgpool-II | General | public | 2016-10-18 23:52 | 2016-12-05 18:07 |
| Reporter | tatzlwurm2 | Assigned To | |||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | open | ||
| Product Version | 3.5.4 | ||||
| Summary | 0000255: Understanding of max_pool | ||||
| Description | http://www.pgpool.net/mediawiki/index.php/Relationship_between_max_pool,_num_init_children,_and_max_connections This says "max_pool*num_init_children <= (max_connections - superuser_reserved_connections) " But if say I have Max_connections=510-10 = 500 Then shouldn't max_pool=5 num_init_children=100 work? What happens is that anything over 100 connections hang. | ||||
| Tags | No tags attached. | ||||
|
|
That's an expected behavior. See: http://www.pgpool.net/docs/latest/pgpool-en.html#NUM_INIT_CHILDREN |
|
|
Thanks but that doesn't answer my question. I even had referred to same link in my description. |
|
|
To be honest, I don't understand your question. If connection clients is over 100 when num_init_children = 100, a "hang" (actually connection request is in the listen queue) is an expected behavior. I don't know what else you are expecting... |
|
|
My question isn't on the num_init_children, its on how max_pool is used. Why have more than one if it can't go over num_init_children. |
|
|
The fomula "max_pool*num_init_children <= (max_connections - superuser_reserved_connections)" is to guarantee for PostgreSQL not to reject connections from Pgpool-II by "sorry, there are too many connections already" error. Not for connections to Pgpool-II from clients. |
|
|
Then how would max_pool>1 be useful? |
|
|
When there are more than 1 user/database pairs. - user a/database b connects to Pgpool-II. It uses 1/max_pool entry to create a cached connection to PostgreSQL. - user c/database d connects to Pgpool-II. If max_pool is 1, it discards the entry for a/b and reuse it. - user a/database b connects again. Since the connection cache for a/b was discarded, it needs to re-connect to PostgreSQL. |
|
|
1 month passed since the last message was added. Issue closed. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2016-10-18 23:52 | tatzlwurm2 | New Issue | |
| 2016-10-19 11:55 | t-ishii | Note Added: 0001124 | |
| 2016-10-21 03:48 | tatzlwurm2 | Note Added: 0001130 | |
| 2016-10-26 16:28 | t-ishii | Note Added: 0001133 | |
| 2016-10-27 03:00 | tatzlwurm2 | Note Added: 0001137 | |
| 2016-10-27 09:33 | t-ishii | Note Added: 0001138 | |
| 2016-11-02 05:34 | tatzlwurm2 | Note Added: 0001148 | |
| 2016-11-02 08:02 | t-ishii | Note Added: 0001149 | |
| 2016-12-05 18:07 | t-ishii | Note Added: 0001199 | |
| 2016-12-05 18:07 | t-ishii | Status | new => closed |
