Best Practices for backing up Cisco Emergency Responder

[tdederscheck]

I have scheduled backups running in DRS however, I have questions about backing up the server as a whole.

1. Does cisco recommend backing up the server that CER is running on?

2. If so does cisco have a supported solution for backing up the server (linux platform)?

Thanks,

Tina

[Anthony Holloway]

DRS is the only supported backup/restore procedure.

I'm curious, what's not good enough about DRS that you wish to backup the entire HDD?  Is it simply, time to restore?

[tdederscheck]

I have seen in several different posts where people say to back up the entire server but I could not find on the cisco site, outside of the discussions, where they supported or promoted the idea.  In general one of the reasons you back up an entire server and not just the data is incase you have to do a restore of the entire box. So I was just curious as to what cisco recommends and what they would support.  It is a linux box and I am still learning the linux platform as well.

[Anthony Holloway]

First, you will not find Cisco saying that you can backup the entire box.  They won't even support pulling a drive out of the mirror as a means of DR.

Second, I still don't understand why you are looking to backup the entire box.  Let's assume worst case scenario, the data center building sinks into a hole in the earth, and you have to rebuild at a new location (don't you have geographically separated clusters anyway?).  You would just order, or pull out of inventory, a new server, pop the install media in, then restore from backup.  Should take you less than 4 hours, once you have the hardware.

[tdederscheck]

Nice.. that is good to know.  The initial question was born from the posts.  It raises the question what exactly does the DRS backup capture?  Does it get configuration changes to the OS such as any changes that had been made to the ip table and so on?

[Anthony Holloway]

I'm not sure what all it grabs.  The docs do not specify, that I saw, but it does state:

...Cisco ER 8.0 web interface, provides full data backup and restore capabilities for all servers in a Cisco ER server group.

Source:http://www.cisco.com/en/US/partner/docs/voice_ip_comm/cer/8_0/English/administration/guide/e911drsc.html#wp1048992

You could easily test if a certain component is backed up by performing a restore on a test box, or as a part of your regularly scheduled DR test procedures.

[William Bell]

I believe that CER DRS operates in the same was as CUCM and CUC. Each node in the cluster will have a set of backup files dumped to your backup server. The backup files are essentially a dozen or so .tar files that contain numerous config files, db info, etc. If you wanted to know what was backed up, you can untar the files and piece it together.

My understanding is that the back up grabs all application settings and database tables. It also grabs host configurations (ip address, hosts files, etc.). To understand what it doesn't grab one should look at the disaster recovery process. To restore the CER DRS backup you must have an operational CER server running the same exact CER version. So, in the event of a disaster you have to build the server from scratch, follow the appropriate install procedures for your version (e.g. Pub first, then sub), then you can perform a restore.

The tricky part is that CER (and other apps as well) use strict version checking. So, you will need to reapply all patches/service releases necessary to get your CER host up to the same level of the server that created the backup file.

This last part is why some folks use methods like breaking the mirror to have an actual backup HDD. Cisco used to condone this practice and actually had tech notes on how to do it. Somewhere along the line Cisco discovered people have a hard time following instruction (sarcasm) and stated they no longer support the practice. At this point in time, I don't subscribe to this practice anymore either.

HTH.

Regards,

Bill