Hi. My predecessor was working on a DR plan that involved a weekly sync between our current server and a DR server at another location. However, he left the complany before this sync was working.
It appears that this sync would end up taking many hours to complete, so I'm wondering if this is an acceptable Disaster Recovery method.
Our other option is a restore from tape.
I'm looking for suggestions or other options for a DR plan.
Any information/suggestions would be appreciated.
Thanks,
Angela
Disaster Recovery Plans
- philmalmaison
- Nouveau
- Posts: 1148
- Joined: 08 Jun 2007 12:00
- Location: Ile de France
you can use control-m server mirroring, having a san system, on twice geographical place, on synchroneous mode
so in this case, it should have fire on the first site, just start failover.
in addition to the concept, do some hot full backup off your database, and be in a transaction log mode to eventually replay near to the crash moment if failover shouldn't work
regards
philmalmaison
so in this case, it should have fire on the first site, just start failover.
in addition to the concept, do some hot full backup off your database, and be in a transaction log mode to eventually replay near to the crash moment if failover shouldn't work
regards
philmalmaison
- th_alejandro
- Nouveau
- Posts: 188
- Joined: 26 Nov 2008 12:00
- Location: Bogotá
DRP Implementation
I, we just implement DRP including Control-M after 1 year of test.
My recomendations :
- Weekly, export xml schemas of your batch window (tables).
- Install binaries from original CD in the DRP machines
- Install your agents in the DRP dependents machines
- Start up Ctm on DRP (0 jobs)
- Upload your xml files
- order your jobs manually (depends on your new day process, if the
order time is so far from new day, many jobs can run at once by
delay time.
- I prefer to install all the component as new installation, that take,
in my case, 3 hours.
What we test ?
We test export database, flash copy of system files, recovery from backups, binary backup, etc. We can't do to work Ctm. At the end, we test an installation from 0, make documentation step by step of the installation procedures and specific configuration for our company. 3 hours for DRP is excellent por us.
My recomendations :
- Weekly, export xml schemas of your batch window (tables).
- Install binaries from original CD in the DRP machines
- Install your agents in the DRP dependents machines
- Start up Ctm on DRP (0 jobs)
- Upload your xml files
- order your jobs manually (depends on your new day process, if the
order time is so far from new day, many jobs can run at once by
delay time.
- I prefer to install all the component as new installation, that take,
in my case, 3 hours.
What we test ?
We test export database, flash copy of system files, recovery from backups, binary backup, etc. We can't do to work Ctm. At the end, we test an installation from 0, make documentation step by step of the installation procedures and specific configuration for our company. 3 hours for DRP is excellent por us.
- th_alejandro
- Nouveau
- Posts: 188
- Joined: 26 Nov 2008 12:00
- Location: Bogotá
another thing....
DRP is Disaster recovery Plan
FailOver is only for specific service
The failover can be implemented in the same location with other available server. Maintain sincronize your CTM database and no problem.
BUt, what appends is no datacenter exist (building crash o great scale damage of ALL the infraestructure in the company ?), that's the solution. DRP.
FailOver is only for specific service
The failover can be implemented in the same location with other available server. Maintain sincronize your CTM database and no problem.
BUt, what appends is no datacenter exist (building crash o great scale damage of ALL the infraestructure in the company ?), that's the solution. DRP.
- philmalmaison
- Nouveau
- Posts: 1148
- Joined: 08 Jun 2007 12:00
- Location: Ile de France
a drp solution could even be mirroring as you can mirror throught out a san system with 2 geographycal different disks implementation.
As all Servers are working in that scope, a multiplex of cluster managment for the applications, and a mirroring system for Control-M Server can allow you to not really need another system to recovory.
regards
philmalmaison
As all Servers are working in that scope, a multiplex of cluster managment for the applications, and a mirroring system for Control-M Server can allow you to not really need another system to recovory.
regards
philmalmaison
Hi...i m newbie. The info which u have shared sounds good. I appreciate with that.
Keep sharing guys...
disaster recovery video
Keep sharing guys...
disaster recovery video
controlm database mirroring
hi,
i just to have the controlm database mirroring setup with a previous client and that always worked as a charm. it is very simple to setup and easy to maintain.
The databases are always up to date and no worries about missing stuff.
i just to have the controlm database mirroring setup with a previous client and that always worked as a charm. it is very simple to setup and easy to maintain.
The databases are always up to date and no worries about missing stuff.