0 votes
690 views
in IBM BPM Installation by

1 Answer

0 votes
by (30.6k points)

A disaster recovery (DR) environment for BPM / BAW is a second, complete cell (or CP4BA cluster) in another data centre that can take over the databases and the traffic. What is installed and configured:

  1. Databases are the heart: BPMDB, PDWDB, CMNDB (and case / content object stores) replicated to the DR site - Db2 HADR (sync / near-sync), Oracle Data Guard, PostgreSQL streaming replication, SQL Server AlwaysOn. The DR databases are standby; failover promotes them.
  2. The DR cell: WebSphere ND + BAW installed identically (same version and fix level, same topology), created with the same BPMConfig properties but pointing at the DR database hostnames - or, better, at DNS aliases that switch with the databases. Cell name, node names and LTPA keys should match production (export / import LTPA keys) so that sessions and SSO behave the same; the messaging engine's file store on shared or replicated storage, or database-backed.
  3. Configuration sync: every change on production (100Custom.xml, resources, certificates, installed snapshots, environment variables) must reach DR - either by replaying the deployment pipeline against DR (recommended: the pipeline installs every snapshot on both) or by periodic profile backups restored on DR. Installed snapshots live in the database, so they arrive with the replication; the managed asset caches and server-specific settings do not.
  4. Front door: the load balancer / GSLB switches the public hostname to the DR IHS pair; certificates present on both sides; LDAP reachable from DR (or a replica).
  5. Process Center: usually not DR-critical (rebuild from twx exports in Git), but a replicated repository database allows the same standby pattern.
  6. Testing: a DR drill twice a year - stop production, promote databases, start the DR cell, run the smoke tests, switch DNS, then fail back; measure RTO (typically 30-60 minutes with warm standby) and RPO (seconds with synchronous replication).
# warm standby: the DR cell exists and is stopped; failover runbook
1. confirm production is down / decision taken       2. promote DR databases (db2 takeover hadr on db BPMDB ...)
3. update DNS aliases for the databases (or nothing if aliases were used)   4. startManager / startNode / startServer on the DR cell
5. verify: login, Process Admin > Event Manager running, a test instance   6. switch the public GSLB / DNS to the DR web tier

On CP4BA: the same idea with a second OpenShift cluster, database replication as above, the CR and secrets in Git (GitOps) applied to the DR cluster, persistent volumes replicated or rebuilt (PFS re-index), and a global load balancer in front of both routers; active-passive is the supported model for the workflow engine (one engine per database at a time).

References

Related questions

0 votes
1 answer 3.0k views
0 votes
2 answers 4.7k views
0 votes
1 answer 1.9k views
0 votes
1 answer 774 views
0 votes
1 answer 882 views
0 votes
1 answer 1.7k views
0 votes
1 answer 2.2k views
0 votes
1 answer 1.4k views
0 votes
1 answer 797 views
0 votes
1 answer 1.3k views
0 votes
1 answer 901 views
0 votes
1 answer 2.2k views
asked Dec 22, 2016 in Interview Questions by Anil Singh (30.6k points)
0 votes
1 answer 750 views

723 questions

807 answers

98 comments

4.9k users

Join BPM Community Discord Channel

Welcome to BPM Tips Q&A, Community wiki/forum where you can ask questions and receive answers from other IBM BPM experts and members of the community. Users with 2000 points will automatically be promoted to expert level.
Created by Dosvak LLC
Our Youtube Channel
...