Disaster recovery? What is the concensus?

[bessebo]

I was just wondering how other Crystal Enterprise administrators deal with disaster recovery? I am backing up all of my reports, instances, and my APS database but would pretty much have to reinstall the application should our main system have a catastrophic failure. I'm wondering how others out there would handle such an occurrence.

Regards,
Bessebo

 

[rhinok]

Sometimes people confuse Disaster Recovery with Fault Tolerance/Failover support.

Disaster Recovery pertains to catastrophic losses, such as your datacenter being destroyed due to a disaster (fire, earthquake, etc...) and your ability to recover from them, sometimes with great effort. For true disaster recovery, you'd have hardware available at an offsite location that could be used in the event of a disaster. It isn't unusual for the recovery to take a matter of days or weeks, depending on your environment. I have heard of environments (post 9/11) that were recovered within hours, however. In regards to Crystal Enterprise, you should definitely have your APS and File Repository Server Objects backed up as often as possible in order to mitigate the risk of losing reports in the event of a failure. You would definitely need to reinstall Crystal Enterprise if it wasn't already loaded on a machine that was waiting.

Fault Tolerance/Failover support generally pertains to hardware and/or software failures and your ability to recover from them, typically with minimal effort. In order to support Failover, you should architect your environment for APS over at least two servers. You should also store the APS/CMS DB and the FRS on servers other than your primary or secondary CE servers. It wouldn't do you any good to set up a clustered environment if the System DB and FRS on the primary server and it goes down...

You still have points of failure (DB Server goes down, for example), but this helps to mitigate risk.


~Kurt

 

[bessebo]

Rhinok,
Thanks for the reply. In regards to security, we have a number of folders and users set up and varying security dependent on the user's functionality. Should the server that Crystal Enterprise is running on have a hard disk failure and we replace the disk and reinstall the application, restore the APS database, and move the File repository objects back into their location, will the security, users, and folders remain intact?

Regards,
Bessebo

 

[rhinok]

You may run into problems if you're using Windows or LDAP security. I'm less familiar with LDAP, but if you're using NT or Active Directory, I'd recommend that you back up the directory structure, or at minimum, keep a list of Windows Users. If you bring up your environment without the mapped users, you run the risk of losing scheduled reports and instances.

Also, I'd recommend that you document all custom configuration (extra commands in the CCM, for example), logins (service account used by various servers) and back up all .csp files (modified for SSO, for example).

Its important to do all these things, but I'd actually recommend that you run your system db and FRS on a different server than your CE server. This way, if the CE server goes down, there's less to recover. If the File/DB server goes down, there's less to recover.

There aren't any guarantees, but if you do all of these things, you have an excellent chance of bringing your environment back up with minimal losses (typically every transaction since the last backup).


~Kurt

 

[Nuffsaid]

Just to add to Kurt’s post, we use CE 10 basically right out of the box. Our system is distributed over 4 dedicated machines. With a complete backup of the CMS database and File Stores, (and with the aid of my “Operating System” buddy) we can go from F-Disk to On Line in about 8 hours.

We've had to do this a couple of times.

Nuffsaid.