Showing results for 
Search instead for 
Did you mean: 

DR solution

Occasional Advisor

DR solution

I have a Debain server ( version 4.0 ) which is running databases and web application , this is critical server which do not allow downtime , I am looking a solution for disaster recovery ( DR ) , can advise is there any option for DR ? does Debian have the feature to do it ? thx
Michal Kapalka (mikap)
Honored Contributor

Re: DR solution


you could use some clustering solutions for debian ant first.

check this :

Frequent Advisor

Re: DR solution

Hi snsd,

for database replication beside classical database replication you can use

DRBD ( )

here is for example for mysql

I setup this on debian and it works very nice.

It works on kernel level and all data are replicated to another host immediately ( I mean you can set up time parameters ). Please read documentation on fist link.

After you carefully set up above ( I suggest first to do all on test system ) you will have database replicated, but still will not failover, that means in case master is down, you will have to manually change ip address at your clients / web server to point to backup.

To avoid this, you can use heartbeat HA
here is link :
all packages exist in debian repository and it is not necessary to install anything else different than .deb packages.

Heartbeat will enable you to have virtual address which will be shared between hosts in communication. So you database will be replicated using DRBD ( or some other method ) and hearbeat will take care to keep alive virtual ip address after master fail and slave will inherit VIP--that means if all configured fine, your applications will not have any problems.
Further if you want to play with loadbalancing then look for ldirector, ipvsadm. Last two will enable you to play with traffic loadbalancing and such nice things.

Hope this helps,

Tim Nelson
Honored Contributor

Re: DR solution

Are you looking for a solution for high availability, e.g. if server is rendered unusable?


Are you looking for DR, e.g. if your facility is rendered unusable ?

The latter is a much bigger topic.

Steven E. Protter
Exalted Contributor

Re: DR solution


DR can be complex or simple. If you can tolerate a days data loss, you can simply replicate a database backup to a DR site once a day.

One thing Linux does not do well is wide area clustering, e.g. one cluster node in your data center and a second in the DR site.

You can, if this is mysql data use a number of data replication strategies that will put a copy of your data, usable at another location, near or far.

Steven E Protter
Owner of ISN Corporation
Valued Contributor

Re: DR solution

Hi All
>One thing Linux does not do well is wide area clustering
it might be true, but I myself build a Oracle10gR2 Standard Edition Cluster/DR using "High Availability Extensions(HAE)" of SUSE Linux Enterprise 10 SP3... and its working rock solid

we have two nodes, the primary/active node is in datacenter, and the other node(passive) is at DR Site( wan clustering)

products/technologies we are using are
1, Heartbeat v2 (for clustering)
2, DRBD for DR(block level replication)
3, we are using 10 MB connection b/w Datacenter and DR Site