Below are ODM DS details that we use
Decision Server 8.9.1.0 , Decision Engine 1.40.5 |
Build #1 on 2017-10-11 14:45:02 Release status: COMMERCIAL |
datasource (Oracle Oracle Database 12c Enterprise Edition Release 12.1.0.2.0 - 64bit Production With the Partitioning, Real Application Clusters, OLAP, Advanced Analytics and Real Application Testing options) |
Wed, 20 Jan 2021 06:16:29 GMT |
Fri, 05 Feb 2021 12:09:24 GMT |
Lately, at the time of RuleApp deployment, we notice that the XOM changes we deploy for a few RuleApps don't sync to one of the two DS nodes we have configured. The node which is gone out of sync still refers to the old XOM (possibly from the cache), resulting in two variety of decisions by HTDS for the same Open API - JSON payload. To overcome this issue, we follow the below practice during the RuleApp deployment using the RES console, and we will make sure the RuleApp is making the correct decisions before we inform the client.
1. Delete XOM
a. XOM: XXXXXXXXXXXXXX-XOM-Domain.zip
b. XOM: XXXXXXXXXXXXXX-XOM-Interface.zip
c. XOM: XXXXXXXXXXXXXX-XOM-Support.zip
2. Delete RuleApp
a. File: XXXXXXXXXXXXXXRuleApp.jar
3. Deploy XOM
a. File: XXXXXXXXXXXXXX-XOM-Domain.zip
b. File: XXXXXXXXXXXXXX-XOM-Interface.zip
c. File: XXXXXXXXXXXXXX-XOM-Support.zip
4. Update Library
a. Library: ADAMgmtRuleApp_1.0
b. XOM Reference: XXXXXXXXXXXXXX-XOM-Domain.zip
c. XOM Reference: XXXXXXXXXXXXXX-XOM-Interface.zip
d. XOM Reference: XXXXXXXXXXXXXX-XOM-Support.zip
5. Deploy RuleApp
a. File: XXXXXXXXXXXXXXRuleApp.jar
b. Option: 'Increment RuleApp major version'
However, yesterday, we noticed the same XOM sync issue for the RuleApp serving fine for a couple of months. We had to do a rolling restart of the nodes to fix the problem.
Any idea what caused the issue to trigger?
Anything on ODM WAS?
Anything that helps to identify the issue using the Execution Unit Dump cache on the RES console?
------------------------------
Charan Paladugu
------------------------------