[ 
https://issues.apache.org/jira/browse/ARTEMIS-2246?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17031525#comment-17031525
 ] 

Per Lundberg commented on ARTEMIS-2246:
---------------------------------------

For the record: what is not mentioned here in this Jira ticket is that the 
actual default in the code [was 
changed|https://github.com/apache/activemq-artemis/pull/2544] in version 2.7.0 
from the previous 100% to 90%. I noticed this after starting to get errors on 
my dev machine. Unfortunately, I had put the log level for 
{{org.apache.activemq.artemis.core.server}} too low in my logback config so it 
took me some time to realize this was the problem...

{{2020-02-06 13:49:54.530 WARN [o.a.a.a.c.server] :: AMQ222210: Free storage 
space is at 13.5GB of 164.3GB total. Usage rate is 91.8% which is beyond the 
configured <max-disk-usage>. System will start blocking producers.}}
{{2020-02-06 13:49:54.931 WARN [o.a.a.a.c.server] :: AMQ222212: Disk Full! 
Blocking message production on address 'media-recording-updates'. Clients will 
report blocked.}}
{{2020-02-06 13:50:04.542 WARN [o.a.a.a.c.client] :: AMQ212054: Destination 
address=media-recording-updates is blocked. If the system is configured to 
block make sure you consume messages on this configuration.}}

The workaround (apart from the obvious "fix" of making more disk space 
available on the partition where the Artemis data is located) is to override 
{{-}}{{Dbrokerconfig.maxDiskUsage=100}} in the Java command line parameters, to 
avoid Artemis blocking until the disk is _really_ full. :)

> Documentation for configuration max-disk-usage is inaccurate.
> -------------------------------------------------------------
>
>                 Key: ARTEMIS-2246
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-2246
>             Project: ActiveMQ Artemis
>          Issue Type: Bug
>    Affects Versions: 2.6.4
>            Reporter: Leo Provido
>            Assignee: Justin Bertram
>            Priority: Major
>             Fix For: 2.7.0
>
>          Time Spent: 2h 50m
>  Remaining Estimate: 0h
>
> During the execution of tests involving sending messages to the System, we 
> observed that the System started blocking messages before the disk is full.
> According to the Configuration Reference 
> ([https://activemq.apache.org/artemis/docs/1.4.0/configuration-index.html]) 
> documentation, the default value of the configuration max-disk-usage is 100. 
> However, we discovered that the actual default value of max-disk-usage is 90. 
> The file inspected is ./etc/broker.xml. Thus, the documentation is deemed 
> inaccurate.
> We have the opinion that severity of this incident is Major because a) the 
> documentation provides an inaccurate piece of information that influences the 
> way we design our mitigation strategies for risks surrounding hardware 
> resources utilization, and b) we don’t necessarily have the bandwidth to 
> review all default configuration values.
> Inaccurate information may cause failure to provide message brokering 
> services despite mitigation strategies being in place.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to