Question
Tata consultancy services limited
US
Last activity: 1 Apr 2016 16:37 EDT
RULES are not hitting
We are seeing sometime now that some of application rules are not getting hit when the flow is executed from the front end application.
We verified the following
1) Verified operator is in correct accessgroup
2) Accessgroup is pointing to current application
3) Application is referring a ruleset which the rules are currently present.
But still we don't see why the particular rule in application rule set is not getting hit.
Pegasystems Inc.
US
Hi Saravanan,
What version of Pega are you using? Can you clarify the behavior that is occurring? Are you seeing rules not being executed at all or are you seeing that a different version is executed?
Carissa
Tata consultancy services limited
US
Hi Carissa-
We are using Pega 7.1. for example the rule is in 02-94-01 version in RSMIAC ruleset. But it is hitting earlier version rule. We see the the RSMIAC: 02-94 is referred in application. No Clue whats happening
Pegasystems Inc.
US
Hi Saravanan,
There could be different reasons for the issue based on the environment and circumstances leading up to the issue.
- What is the Pega ML version?
- What type of database?
- How many nodes in the environment? Is the system pulse running?
- Does the issue only occur for certain types of rules? For example, do you see the issue with stream rules, activities, when rules, etc.
- Does the issue occur when migrating to another environment?
Citigroup technology inc
US
- What is the Pega ML version? 7.2
- What type of database? Oracle DB
- How many nodes in the environment? Is the system pulse running? we have 2 nodes and system pulse running perfectly fine.
- Does the issue only occur for certain types of rules? For example, do you see the issue with stream rules, activities, when rules, etc.
yes it occurs with Activities
- Does the issue occur when migrating to another environment? Yes when we migrate the code from Dev to UAT. We have 2 nodes in UAT. one 1 batch node 4807 and another one is realtime node is 4810. In both the node an activity in the lower version is getting hit when the user invokes a flow from Front end application.
Pegasystems Inc.
US
Hi Saravanan,
I have not come across this issue on Pega 7.2. Does the issue go away if you revalidate and save the rule?
Carissa
PEG
US
If you are using 7.1.7, then everything changes!
Going forward, you should NOT truncate cache tables. I know this has been given as a thing to do in earlier releases, but with the introduction of the VTable cache in the 7 series, except in certain exceptional circumstances, you should not be truncating entire cache tables.
It is likely you have encountered a known error when you did the import (said error not existing in the 7.2 series). In order to avoid future occurrences of this error, please make the following change in your prconfig file; the change will take effect the next time you start the JVM. (Doing this via the DSS path will still require a restart of the JVM to take effect.) This will introduce a 60 second delay between the time you import on one node and the time the system pulse processes the new entry after moving it onto the other nodes. The technical reason why this works is complicated; you will have no performance impact by the change.
Set "systempulse/scanoffsetms" to 60000
In order to fix the problem you currently have, use the SMA to connect to the node on which you have the problem and use the VTable Cache tool (Virtual Rule Table Cache Management) to remove the existing entries for the bad rule from first the database, and then memory (two different buttons). You will need to open the XML of the rule in question in order to provide the required fields -- be careful not to mis-type.
Tata consultancy services limited
US
Thanks for the clarification.
Could you please give a tag to set systempulse/scanoffsetms" to 60000 in prconfig file.
Also in VT Table management under Advanced in SMA, Do we have to give the objclass and instance name of the particular rule to remove the Database and memory cache?
This usually happens when we promote/migrate our code from Dev to UAT. Do we have to always restart the server upon deployment. Because we usually see this getting resolved after restarting the server. But we really not required to restart correct after deployment.
Pegasystems Inc.
US
Here is the format for the prconfig entry.
<env name="systempulse/scanoffsetms" value="60000" />
The virtual rule key uses the format of pxObjClass+colon+pxInsId. Here is an example of what I would enter if I were looking for the AddWork activity.
RULE-OBJ-ACTIVITY:!ADDWORK
You will need to restart you servers after applying the prconfig change so it gets picked up. You should not need to restart your servers are part of a normal rules deployment.
Tata consultancy services limited
US
Thanks again!!
We haven't did these in pega 6 , is there any reason why do we have the prconfig changes in Pega 7.
We already have System pulse running on both the nodes.
PEG
US
Is it the same activity on both nodes that is getting missed? OR is it a different activity on each node?
I'm not familiar with 7.2 to the degree I would like. Can you use the SMA to examine the VTable entries for the activity(activities) of interest and clear both the memory and database VTable entries for same? Clear Database first and memory second, as otherwise the memory might get refreshed from the bad database cache entry. Does the problem then go away?
Tata consultancy services limited
US
Guys sorry we are using PEga 7.1.7. what are the cache tables we should truncate.
We anyway clearing the conclusive cache and memory cache in SMA.