postgresql_cluster
postgresql_cluster copied to clipboard
vitabaks
Consul role uses inside is not support consul versions higher 1.11.x since were introduced breaking changes in config file from consul side
Need to update role according to master branch of role https://github.com/ansible-collections/ansible-consul. Fun starts from version 1.12.x
See https://developer.hashicorp.com/consul/docs/upgrading/upgrade-specific#consul-1-12-x See https://developer.hashicorp.com/consul/docs/upgrading/upgrade-specific#consul-1-13-x See https://developer.hashicorp.com/consul/docs/upgrading/upgrade-specific#consul-1-14-x
tls, addresses ,ports - are affected by these changes master branch of consul ansible role already has it theirs configs
Now we are installing the latest available version of Consul from the hashicorp repository (now it's v1.19) and tests show that it works.
TASK [consul : Looking up latest version of Consul] ****************************
ok: [10.172.0.21]
ok: [10.172.0.20]
ok: [10.172.0.22]
...
TASK [consul : Add hashicorp repository] ***************************************
changed: [10.172.0.22]
changed: [10.172.0.21]
changed: [10.172.0.20]
TASK [consul : Install consul package] *****************************************
changed: [10.172.0.21]
changed: [10.172.0.20]
changed: [10.172.0.22]
...
TASK [consul : Start Consul] ***************************************************
changed: [10.172.0.21]
changed: [10.172.0.20]
changed: [10.172.0.22]
TASK [consul : Check Consul HTTP API (via TCP socket)] *************************
ok: [10.172.0.21]
ok: [10.172.0.20]
ok: [10.172.0.22]
...
TASK [deploy-finish : Postgres Cluster info] ***********************************
ok: [10.172.0.20] => {
"msg": [
"+ Cluster: postgres-cluster (7423961020152107459) --+-----------+-----------------+------------------------------------+-----------------+",
"| Member | Host | Role | State | TL | Lag in MB | Pending restart | Pending restart reason | Tags |",
"+----------+-------------+---------+-----------+----+-----------+-----------------+------------------------------------+-----------------+",
"| pgnode01 | 10.172.0.20 | Leader | running | 2 | | * | max_prepared_transactions: 2000->0 | datacenter: dc1 |",
"| | | | | | | | | key1: value1 |",
"+----------+-------------+---------+-----------+----+-----------+-----------------+------------------------------------+-----------------+",
"| pgnode02 | 10.172.0.21 | Replica | streaming | 2 | 0 | * | max_prepared_transactions: 2000->0 | datacenter: dc1 |",
"| | | | | | | | | key1: value1 |",
"+----------+-------------+---------+-----------+----+-----------+-----------------+------------------------------------+-----------------+",
"| pgnode03 | 10.172.0.22 | Replica | streaming | 2 | 0 | * | max_prepared_transactions: 2000->0 | datacenter: dc1 |",
"| | | | | | | | | key1: value1 |",
"+----------+-------------+---------+-----------+----+-----------+-----------------+------------------------------------+-----------------+"
]
}
TASK [deploy-finish : Connection info] *****************************************
ok: [10.172.0.20] => {
"msg": {
"address": {
"primary": "master.postgres-cluster.service.consul",
"replica": "replica.postgres-cluster.service.consul"
},
"password": "tUnMwGfxt3eZzg15C1VGr3MGqdUqxXnG",
"port": "6432",
"superuser": "postgres"
}
}
I don't use Consul in production (I use etcd), so it's hard for me to say how important these configuration changes really are. Could you take on this task and offer a PR?
I need to check also from my side, since I use this role in customised way. I'll get back a bit later.
yes, and I missed your latest release of this role, ~~so this issue in not relevant since you already points to latest version. But what I'll do it will test this latest role against my consul cluster.~~
Ok, I've checked role config you have here, config.json is not ready for consul higher version than 1.12.x
To fix problem need to use master version of consul role, but master branch has at least 3 opened issue which will potentially impact on cluster deployment.
As example
https://github.com/ansible-collections/ansible-consul/blob/a342eefd6308d92c346324bdcd6506dcf16cd549/templates/config.json.j2#L55-L57
So I'll test and will provide some fixes, after I'll finish with consul cluster upgrade
Started playing with role 2.0
Run this
docker run -d --name pg-console \
--publish 80:80 \
--publish 8080:8080 \
--env PG_CONSOLE_API_URL=http://localhost:8080/api/v1 \
--env PG_CONSOLE_AUTHORIZATION_TOKEN=secret_token \
--env PG_CONSOLE_DOCKER_IMAGE=vitabaks/postgresql_cluster:latest \
--volume console_postgres:/var/lib/postgresql \
--volume /var/run/docker.sock:/var/run/docker.sock \
--volume /tmp/ansible:/tmp/ansible \
--restart=unless-stopped \
vitabaks/postgresql_cluster_console:2.0.0
Login works but everything else no.. )
http://localhost:8080/api/v1/settings?offset=0&limit=999999999 net::ERR_CONNECTION_REFUSED What I missed? @vitabaks
@garry-t Are you running the console locally or on a dedicated server?
If you have set up the console on a different server, replace 'localhost' with the server's address.
And please make sure that ports 80 and 8080 are open.
Ok, thanks. now is ok.
@garry-t Are you running the console locally or on a dedicated server?
If you have set up the console on a different server, replace 'localhost' with the server's address. And please make sure that ports 80 and 8080 are open.
BTW I found that
ok: [10.172.0.20] => {
"msg": {
"address": {
"primary": "master.postgres-cluster.service.consul",
"replica": "replica.postgres-cluster.service.consul"
},
"password": "tUnMwGfxt3eZzg15C1VGr3MGqdUqxXnG",
"port": "6432",
"superuser": "postgres"
}
}
password prints here is not what I set, so far cant figure where this value takes.
patroni_superuser_password
If not defined, will be generated automatically during deployment.
patroni_superuser_passwordIf not defined, will be generated automatically during deployment.
It defined, that is why I'm confusing. But, I'll debug a bit more..