<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=utf-8">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <span id="result_box" class="" lang="en"><span>Hello everyone,<br>
        <br>
      </span><span class="">Currently one of the production gluster
        nodes is consuming a lot of memory, in particular the gluster
        NFS process makes great use of SWAP memory and does not release
        it.</span> <br>
      <br>
      <span class="">It happens to someone else?<br>
        <br>
        # for file in /proc/*/status ; do awk '/VmSwap|Name|Pid/{printf
        $2 " " $3}END{ print ""}' $file; done | sort -k 5 -n -r | less<br>
        <br>
        glusterfs <b>16818</b> 1 0 <b>5283484 kB</b><br>
        glusterfs 16826 1 0 434184 kB<br>
        glusterfsd 14648 1 0 63500 kB<br>
        <br>
        # ps aux | grep 16818<br>
        root     16818 48.1 61.7 27419204 20265168 ?   Ssl  Feb27
        39412:28 /usr/sbin/glusterfs -s localhost --volfile-id
        gluster/nfs -p /var/lib/glusterd/nfs/run/nfs.pid -l
        /var/log/glusterfs/nfs.log -S
        /var/run/gluster/4e5989084b357ad0d2530754e086bd73.socket<br>
        <br>
        # free -m<br>
                               total       used       free     shared   
        buffers     cached<br>
        Mem:         32073      31710        362          0         
        0       4291<br>
        -/+ buffers/cache:      27418       4655 <br>
        Swap:         7247       7247          0 <br>
        <br>
      </span></span><br>
    <span id="result_box" class="" lang="en"><span class=""><span
          id="result_box" class="short_text" lang="en"><span class="">The
            gluster version is<b> 3.7.1-16</b> under RHEL 6.7<br>
            <br>
          </span></span></span></span><br>
    <span id="result_box" class="" lang="en"><span class=""><span
          id="result_box" class="short_text" lang="en"><span class=""><span
              id="result_box" class="" lang="en"><span class="">Launching
                the following commands does not help, as the swap memory
                is re-occupied in full over time.<br>
                <br>
              </span></span># sync<br>
            # echo 3 &gt; /proc/sys/vm/drop_caches<br>
            <br>
          </span></span></span></span><br>
    <span id="result_box" class="" lang="en"><span class=""><span
          id="result_box" class="short_text" lang="en"><span class=""><span
              id="result_box" class="short_text" lang="en"><span
                class="">The swappiness value is currently 60.<br>
                <br>
              </span></span># sysctl -a | grep swapp<br>
            vm.swappiness = 60<br>
            <br>
          </span></span></span></span><span id="result_box"
      class="short_text" lang="en"><span class="">Any idea or similar
        case?<br>
        <br>
        Thanks.<br>
        <br>
      </span></span><span id="result_box" class="short_text" lang="en"><span
        class="">Greetings,<br>
        Jose Antonio Vico Palomino<br>
      </span></span>
  </body>
</html>