Dell DX6000 Getting Started Manual

Also see for DX6000: Manual Supplementary guide Hardware installation instructions Hardware owner's manual Application guide

Copyright © 2010 Caringo, Inc.All rights reserved 16Version 5.0December 2010Amount of RAM Maximum number of immutableunnamed streamsMaximum number of unnamedanchor streams or namedstreams4GB 33 million 16 million8GB 66 million 33 million12GB 132 million 66 million3.4. Stream Size GuidanceThis section provides guidelines you can use to size storage volumes for large stream sizes. Thelargest object a DX Storage cluster can store is one-fifth the size of the largest volume in the cluster.If you attempt to store a larger object, DX Storage logs an error and does not store the object.To further tune your hardware planning, keep in mind that the DX Storage health processor reservesdefragmentation space on a volume equal to twice the size of the largest stream that has beenstored on a volume. Therefore, you might end up having much lower storage capacity than youexpect.If possible, size your hardware so that the largest streams consume between 10 and 20 percentof available space on disk drives used in the storage cluster. If the largest stream consumes 10 to20 percent of disk drive space, you get 60% utilization of available space. The percent utilizationimproves as you add more disk space.For example, if the largest stream consumes between 5 and 10% of disk space, utilization improvesto 80%. If the largest stream consumes only 1.25 to 2.5% of available disk space, utilization is 95%.If disk utilization is diminishing, you should consider upgrading the size of the disk drives in yourcluster nodes.3.5. Adaptive Power ConservationAs of the 4.0 release, DX Storage includes an adaptive power conservation feature thatsupplements DX Storage's naturally green characteristics to spin down disks and reduce CPUutilization after a configurable period of inactivity. A cluster that is constantly in use will likely notbenefit significantly from the adaptive power feature but a cluster that has long periods of inactivityon nights and weekends can expect significant power savings utilizing this feature. Because onlyinactive nodes are affected, max available throughput will not be affected, though additional latencywill be incurred on the first access to a sleeping node. The cluster will automatically awake oneor more nodes to carry out requests when needed and eventually revive all nodes if needed. Theconfiguration parameters that control the adaptive power conservation features can all be set in thenode and/or cluster configuration files.If a node has not serviced any incoming SCSP requests (both client and internode) in the lastconfigurable sleepAfter seconds, it will change to an idle status and pause its health processor,allowing the associated disks to eventually turn idle as well after they have had no IO activityin the past sleepAfter seconds. When an application or another node in the cluster once againbegins sending SCSP requests, one or more of the nodes in the cluster will awake to service thoserequests. Even if no outside activity is detected for a long period, each node will awake after it hasbeen idle for the configurable wakeAfter seconds so that the health processor can monitor disk andcontent integrity.In addition to the sleepAfter and wakeAfter parameters, a new archiveMode setting allows anadministrator to designate a new or empty node as an archive node that will remain idle in low-power mode without participating in cluster activity until its capacity is needed. This allowsadministrators to have additional capacity online and available without paying for the associated