Hi Christopher, Thanks a lot for your information!
As you said, by approach #2 the fail over won't be happened automatically and primary cluster shouldn't be down. So it may not work for us. With approach #1, according to quorum, if we have 3 hosts already in the cluster in primary location, how many host need to be setup in different location. 3+2 will work? when all 3 are down in primary remaining 2 can be fail over? *Or * Do we need 3+3+1, to fail over during DR? Thanks Kalidoss.V On Tue, Jul 26, 2016 at 9:52 PM, Christopher Hamlin <[email protected]> wrote: > For approach 1: > > The cluster needs a quorum: > > > https://help.marklogic.com/knowledgebase/article/View/119/0/start-up-quorum-and-forest-level-failover > > > Approach 2: > > This looks more like DR. Failover will not be automatic. However, if the > replica cluster is running before the master goes down, then it should > continue to run. > > > The documentation for these is good. > > For failover (keep the main cluster highly available), see > > http://docs.marklogic.com/guide/cluster > > For DR (keep another cluster ready), see > > http://docs.marklogic.com/guide/database-replication > > > > > On Tue, Jul 26, 2016 at 6:28 PM, Vinayagamoorthy, Kalidoss < > [email protected]> wrote: > >> >> Hi, >> >> Good Evening!!! >> >> Looking for some help to setup forest replica in a best way to setup >> Disaster Recovery environment. Two approaches have been tried out. Please >> find below the scenarios I have tried and the issues are found against my >> expectations. >> >> >> *Approach #1 - Same Cluster* >> >> Currently, the cluster has 3 hosts in testing environment. The DB is >> split into 3 forests as follows. >> >> - DB-F1 (host1) >> - DB-F2 (host2) >> - DB-F3 (host3) >> >> *DB-F1's* replica setup in host 2&3 >> >> >> - DB-F4-F1(host2) >> - DB-F5-F1(host3) >> >> >> >> *DB-F2's* replica setup in host 1&3 >> >> - DB-F6-F2(host1) >> - DB-F7-F2(host3) >> >> >> *DB-F3's* replica setup in host 1&2 >> >> >> - DB-F8-F3(host1) >> - DB-F9-F3(host2) >> >> >> >> Later we introduced one more host "4" for Disaster Recovery . All split >> forests from host 1,2 & 3 are replicated in host-4 as follows. >> >> DB-F10-F1 - Replica for *DB-F1* >> DB-F11-F2 - Replica for *DB-F2* >> DB-F12-F3 - Replica for *DB-F3* >> >> >> *Test scenario:* >> >> Hosts - 1,2 & 3 are down. Only Host-4 is up. And App service(HTTP) >> should work. >> >> >> *Expectation:* >> >> Host-4 alone is able to handle the data. >> >> *Issues:* >> >> - The fail over his not happened to host-4. >> - Admin page in Host-4 couldn't be loaded. >> - HTTP-API in Host-4 doesn't seem working. >> >> >> >> *Approach #2 - Two Clusters* >> >> In this way, I have created 2 clusters. >> >> - Cluster #1 has Host 1,2 &3. >> - Cluster #2 has Host 4 >> >> >> *Test scenario:* >> >> In cluster #1 all hosts are down >> >> *Expectation:* >> >> Host-4 in cluster #2 alone is able to handle the data. And App >> service(HTTP) should work. >> >> *Issue:* >> >> >> - The fail over is not happened to cluster #2. >> - HTTP-API in Host-4 doesn't seem working. >> >> >> >> Based on my testing both approaches were failed. Looking for the >> suggestion to setup the fail over model in correct way. >> >> >> Thanks >> Kalidoss.V >> >> >> >> >> _______________________________________________ >> General mailing list >> [email protected] >> Manage your subscription at: >> http://developer.marklogic.com/mailman/listinfo/general >> >> > > _______________________________________________ > General mailing list > [email protected] > Manage your subscription at: > http://developer.marklogic.com/mailman/listinfo/general > >
_______________________________________________ General mailing list [email protected] Manage your subscription at: http://developer.marklogic.com/mailman/listinfo/general
