Hello guys,
Is it possible to have 2 different Cognos environments sharing the same content store. the idea is to have a production environment and a replica environment to be used when production environment is down.
Any concerns regarding this approach?
If it is not possible, what is the best approach for such a scenario.
Thanks ..
Seems to be good Idea. I was also waiting if somebody(more experienced) provide suggestion on this. But now thought to share few things as per my knowledge :). I know my comment will not be upto the mark.
Most of the time cognos environment goes down because of the reasons like
some hardware failure at server,
network related issue like some issue with F5 load balancer,
server gets shutdown becuase of some unexpected activity,
there can be issue at authentication source also(LDAP or AD server),
issue on the server hosting content store DB
And most of the time these component will be maintained on the same boxes(for one of the reasons like cost cutting). e.g instead of having three different servers for three different content store DBs(dev,test,prod)..all these DBs will be hosted on the same server. In this case if that server crashes then all three environments will be down. Hence sometimes it becomes difficult to have cognos backup environment, becasue if main(PROD) environment goes down because of some reason at the same time backup environment also can go down becasue of the same reason.
But definitely, idea of sharing the same content store will help in some scenarios.
Take a look at the multi-tenancy support for Cognos 10.2
In here you can have multiple clients / customers using a single content store, there should be no reason why this concept couldn't be applied to different Dev environments.
Content in the store is tagged with a tenant ID so it should be easy to segment content accordingly.
Only downside to this is that you will have to upgrade all environments at the same time if you go for any fix packs or version upgrades.
Cheers
Nick
You can have multiple servers with Cognos components installed and have them point to each other for high availability and load balancing scenarios. We have that here with two servers, but unfortunately I am not the admin and have limited knowledge on how to set it up.
As I understand it you need one content manager running, and a content manager installed and configured on the other boxes, just not set to start automatically. In the event the active content manager machine fails you bring up a dormant content manager, and the other machines use it instead. I am not sure on whether finding the new content manager can be an automatic thing via configuration or not. Perhaps someone more knowledgeable about that can jump in. Some of our databases are similarly configured for high availability, some not, but obviously regardless of the Cognos server state you will need the content database available. So you need to plan for that as well.
Finally note that if you are using VMs in a blade farm, there are also options there to deal with hardware going down and switching to another host, whether local or in another data center.
Hope I have at least provided some food for thought and further investigation.
If you have a standby content manager defined it will automatically become live if the main content manager fails. The switchover isn't seamless, though - because CM also includes CAM, all users need to re-authenticate once the standby content manager comes up.
Cheers!
MF.
That must be how ours is configured, MFGF. I know when there has been an issue with the box normally running our CM everything remains running, but you lose session. I thought maybe that was a lack of automation, but what you say about re-authenticating makes perfect sense. Now my curiosity is up; I'll have to visit with our server dude sometime soon. :)
It is really nice having multiple boxes - two here with a round-robin load balancer in front. When we've had to do something that involves restarting Cognos, he takes down one box at a time and the users seldom notice except for the session issue when the active CM goes offline. Load goes up of course, but most of the time it is not much of an issue from a user perspective. Clearly that could vary depending on your utilization patterns.
Yes dbbits.
As an admin most of the time someone will not stop the services on all servers at a time. Restarting one by one will give the more availability of env. Sometimes when I do only few users will face an issue and as a first try from themselves they will just refresh the page and env will be available for them :).
Admin will follow this one by one approach if the restart of services are required during business hrs.
old post but I think this is a useful thread.
So what would be the best approach in restarting servers in a failover setup?
As I understood, you can restart the main box as long as the standby CM is activated.
And then when the first CM started, does the setup automatically switches back to it as the primary???
When each of your boxes has a Content Manager service, there is no real issue. Basically you only have to restart (Cognos on) the server that has problems. There is no need to restart all Cognos servers. Content Manager will be active on the last failover server.
The question can be answered best when the whole setup of the Cognos environment is known. How many servers, which Cognos components are installed where, are all servers of identical configuration? If not, how are they different?
I was just asking if CM1 comes alive, will it be automatically switch as the active CM again.
Cheers!
Quote from: sdf on 06 Aug 2018 07:50:17 AM
I was just asking if CM1 comes alive, will it be automatically switch as the active CM again.
Cheers!
Hi,
No. If the standby CM is active, the main CM will go into standby mode when it restarts.
MF.
yep!
thanks MF!