[jboss-jira] [Red Hat JIRA] (WFLY-14275) Large job repository is blocking deployment
[
https://issues.redhat.com/browse/WFLY-14275?page=com.atlassian.jira.plugi...
]
Cheng Fang commented on WFLY-14275:
-----------------------------------
To optimize it with one single query will need changes to jberet-core (as outlined in
JBERET-506), which will offer a long-term fix.
As a partial fix directly in WildFly, I propose the following improvements to
https://github.com/wildfly/wildfly/blob/master/batch-jberet/src/main/java...
1, limit the max number of job instances (e.g., 50) to retrieve during refreshing
process;
2, avoid excessive refreshing, by keeping track of the last refreshing time, and only
refresh after certain interval (e.g., 1000 ms) elapses.
Large job repository is blocking deployment
-------------------------------------------
Key: WFLY-14275
URL: https://issues.redhat.com/browse/WFLY-14275
Project: WildFly
Issue Type: Bug
Components: Batch
Reporter: Gunther vw
Assignee: Cheng Fang
Priority: Critical
Attachments: jberet-deploy.zip
We are updating from WildFly 13 (jberet 1.2.4) to WildFly 21.0.2 (jberet 1.3.7).
In our production environment we are running about 50 jobs, some of them in a frequency
of 2 or 4 minutes. For documentation reasons we only can delete job information older than
3 months. So we have 20000 and more jobinstances per job in our job repository - an oracle
database job repo.
Deploying of our Job-Module to WildFly 21 is never finished - sometimes a out of memory
exception is thrown.
--
This message was sent by Atlassian Jira
(v8.13.1#813001)