Incorrect disk ("fs") stats with multiple nodes in a single box #24472
Description
Elasticsearch version: 5.3.0
Plugins installed: [none]
JVM version: Java 1.8
OS version: Windows (but is happening as well in Linux)
Description of the problem including expected versus actual behavior:
Have two nodes in Disk D, one pointing to a data folder in Disk G (node1) and the other to Disk F (node2) both with 2GB each:
Disks:
Nodes stats clearly shows that we have 2 nodes, each with 2GB:
{ "nodes" : { "VEpm8eR8SjmV0AEXOy3DUA" : { "fs" : { "timestamp" : 1493837976833, "total" : { "total_in_bytes" : 2144333824, "free_in_bytes" : 2102579200, "available_in_bytes" : 2102579200 }, "data" : [ { "path" : "G:\\data\\nodes\\0", "mount" : "Node1 (G:)", "type" : "NTFS", "total_in_bytes" : 2144333824, "free_in_bytes" : 2102579200, "available_in_bytes" : 2102579200 } ] } }, "ItZtzg8ER-WBvMUo2eEYvA" : { "fs" : { "timestamp" : 1493837976836, "total" : { "total_in_bytes" : 2144333824, "free_in_bytes" : 2102579200, "available_in_bytes" : 2102579200 }, "data" : [ { "path" : "F:\\data\\nodes\\0", "mount" : "Node2 (F:)", "type" : "NTFS", "total_in_bytes" : 2144333824, "free_in_bytes" : 2102579200, "available_in_bytes" : 2102579200 } ] } } } } But the cluster stats are showing only half of that:
{ "nodes" : { "fs" : { "total_in_bytes" : 2144333824, "free_in_bytes" : 2104438784, "available_in_bytes" : 2104438784 } } } Behavior: total shows 2GB
Expected: total should show 4GB
Steps to reproduce:
- Start two nodes one pointing to a mount, the other to another mount
- Run
http://localhost:9200/_cluster/stats?pretty&filter_path=nodes.fs - Verify that total in bytes is 2GB, while it should be 4GB.
This is affecting also Monitoring, as it shows wrong information:
