You are here

What do you do for resilience?

Submitted by cliffordmarks on Mon, 12/17/2012 - 07:24

For most university institutions if not all it is probably not possible to have automatic fail over or luminis 4 portal but do any universities have a semi automatic fail over strategy?

What do you do for resilience?
- I.e. Do you have a replicated portal on another site within your campus or out?
- Is it a full exact replication of your live setup or a reduced setup.
- Do you also have a replication of dev and test environments also?

Luminis Version:

Do you have a replicated portal on another site within your campus or out?

- Yes. It is identical to our Production environment running parallel deployment. the only difference is this failover environment runs on 8 webservers and our prod runs on 15.

Is it a full exact replication of your live setup or a reduced setup.

- See above. We have a failover Banner environment too and the Luminis failover environment points to that one since Luminis and Banner have a 1:1 relationship.

- Do you also have a replication of dev and test environments also?

- No.

Hi
Thanks for your response. We are just in the process of going through another resilience project and reviewing our current resilience setup.

If a disaster happened that destroyed your current Production environment:
- how quickly have you estimated that it would take to get your failover environment up and running?

- How do you failover? via a loadbalancer balancing between production and failover setups? or manually bringing down network interfaces etc...?

- have you ever had to failover? Or do you periodically go through a resilience check/test run?

Thanks

Hello Vikram,
We are in the process of setting up a failover environment with our luminis 4.3.

Do we need to setup luminis database in a data guard environment, Is ldap setup as a consumer to the primary ldap or as a multi master setup will be good enough to capture all the changes made from the web server/portal?
Can anyone share your approach in detail
Thank you for the time.

We are working on a second data center. The plan is to have our test portal in the second data center, and fail over to that if prod is down.

The server guys are working on a vmware solution that would periodically copy the prod virtual servers over to a storage location at the test site. Those prod copies would be brought up on the test hardware when needed.

The test systems ideally would match production, but based on history, we will likely end up with a scaled down version in terms of hardware at the test site.

They are planning similar fail over strategies for Banner and other critical systems.

All,
We are in the process of setting up a failover environment with our luminis 4.3.

Do we need to setup luminis database in a data guard environment, Is ldap setup as a consumer to the primary ldap or as a multi master setup will be good enough to capture all the changes made from the web server/portal?
Can anyone share their approach in detail
Thank you for the time.

You shouldn't need to do any product (ldap,oracle) specific HA stuff if you have things virtualized in vmware. The vmware images can be live sync'd to another data center (similar to rsync), and turned on if the primary dies.

thank you for your reply, at this point we are not planning to replicate at the vmware level but more so at the application.
I have the complete luminis configureed environment and trying out different options like.

will syncing only ldap be enough, or do I need to also sync uportal schemas?