[ARTEMIS-872] On extremely large volumes, Artemis falsely reports "Storage usage is beyond max-disk-usage"

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.5.0
• Fix Version/s: 2.5.0
• Component/s: Broker
• External issue URL: https://issues.apache.org/jira/browse/AMQ-6441
• Assignee: Justin Bertram
• Reporter: Laurent Gauthier
• Created: 06/Dec/16 00:07
• Updated: 21/Feb/18 13:22
• Resolved: 21/Feb/18 13:22

Description

On Amazon EFS (Elastic File System) Java 8 reports the following for UsableSpace and TotalSpace:

usable=9,223,372,036,635,623,424
total= -9,223,372,036,854,775,808

As one can see, the totalSpace is negative! We are past the Long.MAX_VALUE. This results in org.apache.activemq.artemis.core.server.files.FileStoreMonitor reporting an error of type "Storage usage is beyond max-disk-usage" since it computes usage (in method calculateUsage(FileStore store)) as:

1.0 - (double) store.getUsableSpace() / (double) store.getTotalSpace();

so that "usage > maxUsage" resolves to True (maxUsage = 0.999 by default).

This problem was already reported and discussed for ActiveMQ 5.x (see External Issue URL) and a proposed fix was to use the following logic:

if (totalSpace < 0)

{ totalSpace = Long.MAX_VALUE; }

The other proposal is to support a flag that would disable the usage check altogether.

Issue Links

• relates to: AMQ-6441 Incorrect File System Size Reported with Amazon Elastic File System (EFS)
• links to: GitHub Pull Request #1878