AnsweredAssumed Answered

inode limit exceeded

Question asked by stormcrow on Dec 8, 2013
Latest reply on Dec 13, 2013 by Ted Dunning
We're getting an inode limit exceeded warning from the cluster for our root volume. I started investigating:

<code>
 # maprcli dump volumeinfo -volumename mapr.cluster.root -json
</code>

This spits out a lot of information, but what really gets my attention is this json:

                    {
                            "ContainerId":2049,
                            "Epoch":68,
                            "Master":"192.168.3.148:5660--68-VALID",
                            "ActiveServers":{
                                    "IP:Port":[
    "192.168.3.148:5660--68-VALID",
    "192.168.3.147:5660--68-VALID"
                                    ]
                            },
                            "InactiveServers":{
    
                            },
                            "UnusedServers":{
    
                            },
                            "OwnedSizeMB":"131.57 GB",
                            "SharedSizeMB":"0 MB",
                            "LogicalSizeMB":"131.57 GB",
                            "TotalSizeMB":"131.57 GB",
                            "NumInodesInUse":20141241,
                            "Mtime":"Sun Dec 08 09:48:42 EST 2013",
                            "NameContainer":"true"
                    },

Every other container is listing a number of inodes in use on the order of a few hundred. This one is listing 20 million. It's also a whopping 130 GB in size, while the rest of the containers are 15 GB. Is this indicative of something in particular? How should I proceed?

Outcomes