[ 
https://issues.apache.org/jira/browse/ARTEMIS-4065?focusedWorklogId=832832&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-832832
 ]

ASF GitHub Bot logged work on ARTEMIS-4065:
-------------------------------------------

                Author: ASF GitHub Bot
            Created on: 12/Dec/22 17:50
            Start Date: 12/Dec/22 17:50
    Worklog Time Spent: 10m 
      Work Description: clebertsuconic commented on code in PR #4265:
URL: https://github.com/apache/activemq-artemis/pull/4265#discussion_r1046173412


##########
artemis-commons/src/main/java/org/apache/activemq/artemis/utils/SizeAwareMetric.java:
##########
@@ -163,6 +173,21 @@ public final long addSize(final int delta) {
       return addSize(delta, false);
    }
 
+   public final void add(final int elements, final long size) {
+
+      long currentSize = sizeUpdater.addAndGet(this, size);
+      long currentElements = elementsUpdater.addAndGet(this, elements);
+
+      if (elements >= 0) {
+         assert size >= 0 : "If elements is positve, size must be positive";

Review Comment:
   @gemmellr thanks for the catch!





Issue Time Tracking
-------------------

    Worklog Id:     (was: 832832)
    Time Spent: 1h 20m  (was: 1h 10m)

> Option to use non persistent counters in paging. Rebuild them upon start if 
> persistence is disabled on them
> -----------------------------------------------------------------------------------------------------------
>
>                 Key: ARTEMIS-4065
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-4065
>             Project: ActiveMQ Artemis
>          Issue Type: Improvement
>            Reporter: Clebert Suconic
>            Priority: Major
>          Time Spent: 1h 20m
>  Remaining Estimate: 0h
>
> Instead of storing records on journal for counting how many records there are 
> in paging, the system should instead just swipe the paging system in parallel 
> with processing data.
> The changes I'm making will take a snapshot of the current records of paging, 
> and then it will read all the pages to rebuild the counters.
> On tests I am making from a real data server, a system that had a lot of 
> pages (700) needed less than 1 minute to rebuild the counters, and the 
> messages were available to be delivered while the swipe was being done.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to