Description
Since Zowe 3.0, multiple companies including mine have ran Zowe only to find this happen:
2025-02-24 18:08:11.844 <ZWEAGW1:DiscoveryClient-InstanceInfoReplicator-%d:67568771> me WARN ((o.s.b.a.s.DiskSpaceHealthIndicator)) Free disk space at path '/runtime/zowe-release-3.1.0/components/gateway/.' below threshold. Available: 8249344 bytes (threshold: 10485760B)
2025-02-24 18:08:12.845 <ZWEAGW1:DiscoveryClient-InstanceInfoReplicator-%d:67568771> me WARN ((o.s.b.a.s.DiskSpaceHealthIndicator)) Free disk space at path '/runtime/components/gateway/.' below threshold. Available: 8249344 bytes (threshold: 10485760B)
That is, the gateway server spams the log every second that your runtime directory is perceived to be below 10MB free.
There's several issues with this, but...
- It shouldn't care, because it shouldn't be writing to the runtime directory
- It's not even true. z/OS surprisingly does not implement
df
properly, or whatever it's underlying API is, and it doesn't really show free space at all. It might be showing free space of CURRENT EXTENT, but it's easy to see free space INCREASE by copying a big file in.
This behavior seems to be the default behavior of https://docs.spring.io/spring-boot/api/java/org/springframework/boot/actuate/system/DiskSpaceHealthIndicator.html wherever it is being used.
Not sure why spring does this.
It's suggested elsewhere that this behavior can either be disabled broadly with management.health.diskspace.enabled=false
or moved to somewhere beneficial like the workspace with management.health.diskspace.path=/new/path/
But, considering that it spams each second, I don't think pointing it at a different directory is 'good', disabling it would be better.
Metadata
Metadata
Assignees
Type
Projects
Status