Hi,
1. One manager means that you are not highly available, If this host goes down there is no recovery which is not that good in production unless you have all your data stored in database as well and recovery from the database takes a time that you can afford.
Usually in production we recommend 3 managers to allow quorum, Anyhow if you had the same on host with 9.5 version you are aware of the problems in working with just one host.
2. Moving to manager will allow you to use Rest API, new CLI and ops manager (manager_ip:8090\v2, Manager_ip:8090) so its recommended even when you have just one machine
3. To start each gsc with its zone :
To start a manager set GS_MANAGER_SERVERS in setenv-overrides script
Step 1: Make sure to start a manager
./gs.sh host run-agent --manager
Step 2: Use gs.sh container option *
./gs.sh container create --zone=zone1 --memory=1g <hostname>
Same can be also achieved with REST API
Regards,
Ester
Cezary Kowalczuk
Hi!
We're migrating from xap 9.5 to 15.2. Our application is hosted on 1 machine. We need 2 gsc, one for processing unit (jar) and second for web application (war).
What's the best option for deployment for this case?
We've been using scripts to deploy, and it was covering all of our needs. But now, according that it will be deprecated in next versions, we want to try Manager.
Out grids were started by this (ommited variables export):
1. management zone:
gs-agent.sh gsa.gsc 0 gsa.global.gsm 2 gsa.global.lus 2
2. processing unit zone:
gs-agent.sh gsa.gsc 1 gsa.global.gsm 0 gsa.global.lus
3. web zone
gs-agent.sh gsa.gsc 1 gsa.global.gsm 0 gsa.global.lus
What's the best equivalent for this structure to use manager? And is this okay option if we use only 1 host for prod?
Thanks