Phil - Gcollect is agnostic as to the data collector, it is designed to support adding collectors on the fly. Using the collecd-exec plugin and PUTVAL it should be easyish to add collectd support. Here is (roughly) what you need to do - copy ganglia.d/* to collectd.d/* edit the bottom of each file in collectd.d, basically you need to replace the gmetric commands with PUTVALs - gmetric --name=brick_${GLUSTER_BRICK_MP}_disk_read_io --value=${IO_KBS_READ} --type=int32 --units=kilobytes --group=${GLUSTER_VOLNAME} gmetric --name=brick_${GLUSTER_BRICK_MP}_disk_write_io --value=${IO_KBS_WRITE} --type=int32 --units=kilobytes --group=${GLUSTER_VOLNAME} gmetric --name=fuse_net_io_rx --value=${NET_IO_KBS_RX} --type=int32 --units=kilobytes --group=${GLUSTER_VOLNAME} gmetric --name=fuse_net_io_tx --value=${NET_IO_KBS_TX} --type=int32 --units=kilobytes --group=${GLUSTER_VOLNAME} gmetric --name=nfs_net_io_rx --value=${NFS_IO_KBS_RX} --type=int32 --units=kilobytes --group=nfs-common gmetric --name=nfs_net_io_tx --value=${NFS_IO_KBS_TX} --type=int32 --units=kilobytes --group=nfs-common would become something like - PUTVAL $HOSTNAME/$GLUSTER_VOLNAME/$GLUSTER_BRICK_MP/disk_read_io interval=600 N:$IO_KBS_READ PUTVAL $HOSTNAME/$GLUSTER_VOLNAME/$GLUSTER_BRICK_MP/disk_write_io interval=600 N:$IO_KBS_WRITE PUTVAL $HOSTNAME/$GLUSTER_VOLNAME/fuse_net_io_rx interval=600 N:$NET_IO_KBS_RX PUTVAL $HOSTNAME/$GLUSTER_VOLNAME/fuse_net_io_tx interval=600 N:$NET_IO_KBS_TX PUTVAL $HOSTNAME/$GLUSTER_VOLNAME/nfs_net_io_rx interval=600 N:$NFS_IO_KBS_RX PUTVAL $HOSTNAME/$GLUSTER_VOLNAME/nfs_net_io_tx interval=600 N:$NFS_IO_KBS_TX Change line 9 of gcollect.cfg to "collector = collectd" Then you would want to execute gcollect using collectd-exec every 5 min. Does that help you get started? Craig craig at gestas.net