You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "J. Ryan Earl (JIRA)" <ji...@apache.org> on 2013/10/10 16:47:42 UTC

[jira] [Created] (CASSANDRA-6179) "Load" calculated in "nodetool info" is strange/inaccurate in JBOD setups

J. Ryan Earl created CASSANDRA-6179:
---------------------------------------

             Summary: "Load" calculated in "nodetool info" is strange/inaccurate in JBOD setups
                 Key: CASSANDRA-6179
                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6179
             Project: Cassandra
          Issue Type: Bug
         Environment: JBOD disk layouts
            Reporter: J. Ryan Earl


We recently noticed that the storage capacity on Cassandra nodes using JBOD layout was returning what looks close to the average data volume size, instead of the sum of all JBOD data volumes.  It's not exactly an average and I haven't had time to dig into the code to see what it's really doing, it's like some sort of sample of the JBOD volumes sizes.

So looking at the JBOD volumes we see:
{noformat}
[jre@cassandra2 ~]$ df -h
Filesystem            Size  Used Avail Use% Mounted on
[...]
/dev/sdc1             1.1T  9.4G  1.1T   1% /data/1
/dev/sdd1             1.1T  9.2G  1.1T   1% /data/2
/dev/sde1             1.1T   11G  1.1T   1% /data/3
/dev/sdf1             1.1T   11G  1.1T   1% /data/4
/dev/sdg1             1.1T  9.2G  1.1T   1% /data/5
/dev/sdh1             1.1T   11G  1.1T   1% /data/6
/dev/sdi1             1.1T  9.8G  1.1T   1% /data/7
{noformat}

Looking at 'nodetool info' we see:
{noformat}
[jre@cassandra2 ~]$ nodetool info
Token            : (invoke with -T/--tokens to see all 256 tokens)
ID               : 631f0be3-ce52-4eb9-b48b-069fbfdf0a97
Gossip active    : true
Thrift active    : true
Native Transport active: true
Load             : 10.57 GB
{noformat}

So there are 7 disks in a JBOD configuration in this example, the sum should be closer to 70G for each node.  Maybe we're misinterpreting what this value should be, but things like OpsCenter appear to use this "load" value as the size of data on the local node, which I expect to be the sum of JBOD volumes.



--
This message was sent by Atlassian JIRA
(v6.1#6144)