Showing results for 
Search instead for 
Did you mean: 

ePO Disaster Recovery

Hi All,

Maybe slightly off the thread topic but the process to recover after an ePO Mater server fails in some way is one that interests me.

Is there documentation or guides that describes the steps to take in the event of this?

In our environment we have the following in place

4 hour response for Hardware support
Take Full backups to external media

However, if we lost a site what are the steps needed to rebuild the ePO Master server. During this downtime are there any workarounds for getting Polcies say for HIP's 7.0 applied to the client endpoints?

Sorry this is very generic and slighlty off topic.

4 Replies

RE: ePO Disaster Recovery

I have splitted your post to a new topic.

Regarding you question a recovery of the epo server (master repository) can be done the following way:

1) Reinstall OS
2) Reinstall ePO 4 with the exact same patch level and exact same port configuration
3) Checkin all packages and extensions you had before
4) Restore you security keys (be sure to EXPORT your security keys from your existing server regularly)
5) Delete new initial database and restore backup database

Regarding policies, thr clients will only get them from your ePO Server, so if it is down there is no way to apply new policies.


Thanks Tom..again.

I am now starting to get concerned that this is the only option for disastery recovery and this now links back to my initial post about the ePO design where you have suggested using only one ePO server and one distributed repository. I am aware that the solution we are outlining is limited due to limited chassis and indeed not taking advantage of clustering.

Sorry to be a pain but could you please outline why you have suggested using only one ePO server and not both our severs being master ePO servers? Is it due to being technically difficult\impossible or another reason altogther?

In my defence...I have had only a few days to get to grips with the product Smiley Happy


RE: Update

Asking is no problem :-)

The only option to get a real failover solution is to install a epo server cluster.
Installing two seperate epo servers will not be a failover because (currently) the agent can only access one epo server. For feature releases it is planned to configure more epo servers for one agent.

epO agent


It now becomes clearer, I think I'm getting there and should have realised that as I read it yesterday

When the FRAMEPKG.EXE is created it contains the IP address of the server.....etc.