I am in the process of setting up Informix HDR on informix 11.
I need your inputs and assistance. Can you please give me the steps to configure HDR. Also would like to know whether the main requirement for HDR is two servers with identical configuration or can we setup HDR in two instances on same server, which can be don in ER. Also is it possible to run both HDR and ER on same server. Appreciate your help.
HDR is pretty straight forward but is slightly easier to explain when using two seperate machines. Before I go through the steps you must ensure both machines are the same OS, same processor and disk capacity. Memory is not so much an issue.
1. Install IDS 11 on both machines.
2. Configure onconfig on both machines - some settings need to be the same but those covering memory do not.
3. Configure sqlhosts on both servers AND ensure both servers trust each other. On Unix this can be achieved with /etc/hosts.equiv - one windows %ROOT%\sytem32\drivers\etc\hosts.equiv. The rule is you need to be able to access the instance on the other machine without entering a username and password.
4. Initialise the primary server. If the primary is already up and running that's fine.
5. Archive the primary - I use ontape so for me the command is "ontape -s -L 0"
6. On the secondary, the disk layout needs to be prepared. If using explicit (rather than relative) paths for all chunks, the chunk files need to be in the same folder with the same name and permissions.
7. Restore the tape on the secondary, again using ontape this would be "ontape -p"
8. Issue the pairing commands:
On the primary "onmode -d primary ol_nameofthesecondaryinstance"
On the secondary "onmode -d secondary ol_nameoftheprimaryinstance"
Keep an eye on both servers with onstat -m. Any issues will be reported there but hopefully eventually you will see "Primary Server Operational" in the primary's log.
To get HDR working on the same box you have to ensure the primary and secondary chunks can be created in seperate folders. This can be achieved by initialising the primary chunks with relative paths i.e. instead of a chunk being created with an explicit path e.g. /data/base/rootdbs they are relative e.g. ./rootdbs.
You will need to script the setting up of your environment so you can switch between instances by defining variables like DBSERVERNAME, ONCONFIG easily.
ER is a completely different kettle of frogs in terms of defining the replicated objects but as you can have multiple instances on the same machine it is possible.
I have a single test machine but have multiple instances with combinations of HDR/RSS and the occasional ER server.