Clariion Timeline Clariion Timeline All members of the CX family have a similar architecture. The main differences are the number of front-end and back-end ports, the CPU types and speeds, and the amount of memory per SP. Clariion Hardware
Clarion Hardware Architecture Delivering Data and Application Availability Fully redundant architecture SP, cooling, data paths, SPS Non-stop operation Online software upgrades Online hardware changes
No single points of failure Tiered capacity FC and ATA disks From five to 480 disks
Continuous diagnostics Data and system CLARalert Flexibility Mix drive types, RAID levels RAID levels 0, 1, 1+0, 3, 5
Advanced data integrity Mirrored write cache De-stage write cache to DISK upon power failure Up to 16 GB memory
Dual I/O paths with no disruptive failover Clariion Architecture Clariion Architecture is based on intelligent Storage Processors that manage physical drives on the back-end and service host requests on the front-end. Depending on the module, each Storage processor includes either one or two CPUs. Storage Processors
communicate to each other over the CLARiiON Messaging Interface (CMI). Both the front-end connection to the host and the back-end connection to the physical storage are 2Gb/4GB Fibre channel CLARIION Features Data Integrity How CLARiiON keeps data safe (Mirrored write cache ,vault, etc)
Data Availability Ensuring uninterrupted host access to data (Hardware redundancy,pathfailover software(powerpath), Error reporting capability) CLARiiON Performance What makes CLARiiON a great performer
(cache, Dual SPs , Dual/quad back-end FC buses ) CLARiiON Storage Objects A first look at LUNs, and access to them ( RAID Groups, LUNs , MetaLUNs,Storage Groups) Modular Building Blocks in Storage system The CLARiiON storage system is based upon a modular architecture. There are four building blocks in a Clariion.
DPE - Disk Processor Enclosure Contains both disks and processor DAE - Disk Array Enclosure Contains disks only SPE - Storage Processor Enclosure Contains storage processor
SPS - Standby Power Supply Provide battery backup protection The DPE houses the storage Processor(s) and the first set of Fibre Channel disks. The DPE includes: Two power supplies ,each have a power input connector that is fed by SPS Two Storage Processors that include the SP and LCC functionality. Each SP has memory and one or more processors. Backend ports , Frontend Ports ,Serial port and Ethernet Management port
DAE (Disk Array Enclosure) Disk Status LEDs Green for connectivity Blinks during disk activity Amber for Fault Enclosure Status LEDs Green = Power Amber = Fault
DAE (Disk Array Enclosure) DAE SPA (Storage Processor Enclosure) Front View of SPA SPE Rear view of SPE SPS (Standby Power Supplies)
The CLARiiON is powered on or off using the switch on the SPS. The RJ11 connection is to the Storage processor and used to communicate lost of AC power and signals the SP to begin the vault operation. Once the vault operation is complete, the SP signals the SPS that it is OK to remove AC power Note: Until the batteries are fully charged, write caching will be disabled DAE-OS Front view The DAE-OS contains slots for 15 dual-ported Fibre Channel disk drives. The first five drives are referred to as the Vault drives.
Disks 0-3 required to boot the Storage Processors Disks 0-4 required to enable write caching These disks must remain in the original slots! The DAE-OS enclosure must be connected to bus zero and assigned the enclosure address 0. Private Space Private space on Vault/Code Drives The first five drives in DAE are called code drives They are also used for vaulting purpose.
6.5 GB of each drive of code drives is reserved to store Flare image, SPA and SPB boot images and for PSM LUN and Vaulting purpose Flare is triple mirrored PSM LUN triple mirrored Vault: Vault is a reserved area found on 1 st nine disks of DPE in FC series and 1 st five disks of DPE on CX series. Data in write cache is dumped to the vault area in power failure emergency. Once the system is turned on vault transfers dumped data back to cache PSM LUN: Persistent Storage Manager LUN ,created at the time of initialization by Navisphere
PSM is a hidden LUN where the records of configuration information and access logix database are stored. It resides in the first three disks of code drives Both SPs can access a single PSM and update themselves with new configurations via Clariion Messaging interface(CMI) Clariion Operating Environment The CLARiiON arrays boot operating system is either Windows NT or Windows XP depending on the processor model
After booting each SP Executes FLARE software. FLARE software manages all functions of the CLARiiON storage system(provisioning, resource allocation, Memory management etc. Access Logix software is optional software that runs within the FLARE operating environment on each storage processor (SP).It is used for LUN masking Navisphere provides a centralized tool to monitor, configure, and analyze performance of clariion storage systems. CLARiiON can also be managed as part of EMC ControlCenter, allowing full end-to-end management. Other array software includes SnapView, MirrorView, and SANCopy.
Clariion Management Software Components Software Components Array Software Base (FLARE) code (with or without Access Logix) Array agent Management Server
Management UI SnapView MirrorView Management Station Software Internet Explorer or Netscape Java
Navisphere Management UI ClarAlert Host Software Navisphere Host Agent HBA drivers PowerPath Note: The Navisphere UI may run either on the management station or on the array. Navisphere Users
There are three roles of users: Administrator Can do anything including create and delete users. Manager Can fully manage array but cannot modify/create/delete other users. Monitor Can only look. There are two scopes: Local Global Clariion configuration
Introduction to Navisphere Manager Configure the Clariion Clarion Security ( Domain configuration and Creaing user A/Cs etc Configure Cache, Verify available softwares, acess logix, Network configuration, Verify SPs WWNs and setting SP agent privileged users etc) Create RAID groups
Bind LUNS and MetaLUNs Initiator Records and host registration Access logix
Create storage groups RAID groups and LUNS RAID Group: RAID Group is a collection of Physical Drives from which an administrator may bind one or more LUNs. Once the first LUN is bound within a RAID group, all other LUNs will the RAID Group will share the same protection scheme.
Using the Navisphere GUI and or CLI we can administer RAID groups(Create, Expand, Destroy etc) LUN: LUN is a Logical Unit The process of creating a LUN is called Binding When presented to a host it is assigned a Logical Unit Number and it appears to the host as a disk drive
Using the Navisphere GUI and or CLI we can administer LUNs( Bind LUN, Changing LUN properties, Unbinding LUN etc) RAID groups and LUNs MetaLUN: Collection of individual LUNs that act in together with, and are presented to, a host or application as a single storage entity Created by taking new and/or pre-existing
LUNs and logically connecting them together Expand existing volumes while on-line Concatenated Striped Combined Stripe and Concatenated
MetaLUN Terminology FLARE LUN (FLU) A logical partition of a RAID group. The basic logical units managed by FLARE, which serve as the building blocks for MetaLUN components. MetaLUN A storage volume consisting of two or more FLUs whose capacity grows dynamically by adding FLUs to it Component A group of one or more FLARE LUNs that get concatenated to a MetaLUN as a single or striped unit Base LUN
The original FLARE LUN from which the MetaLUN is created. The MetaLUN is created by virtue of expanding the base LUNs capacity. Note : The MetaLUN is presented to the host in exactly the same way it was before the expansion i.e. the Name, LUN ID, SCIS ID, and WWN is the same. The only thing that changed is the capacity is increased. To Expand a LUN, right click on the LUN and select Expand This invokes the Storage Wizard LUN 0
LUN Mapping SPA t0 t1 S t2 W t3 I T SPB C
H HOST Access Logix Access Logix
What Access Logix is Why Access Logix is needed Configuring Access Logix Storage Groups Configuring Storage Groups Access Logix
Access Logix is a licensed software package that runs on each storage processor. SAN switches allow multiple hosts physical access to the same SP ports . Without Access Logix, all hosts would see all LUNs.
Access logix solve this problem using LUN Masking by creating Storage groups. Controls which host have access to which LUNs Allows multiple hosts to effectively share a CLARiiON array Manual and Auto Registration Automatic Registration: Registration is performed automatically when a HBA is connected to an array There are two parts to the registration process:
Fibre Channel port login (plogi) where the HBA logs into the SP port Creates initiator records for each connection Viewed in Navisphere in Connectivity Status Host Agent registration where the host agent completes the initiator record information with host information Manual Registration: The Group Edit button, on the Connectivity Status main screen, allows manual registration of a host which is logged in to. In FC series we need to do manual registration. In CX series the registration is done
automatically if Host agent is installed on Fabric hosts Storage Groups Managing Storage Groups
Creating Storage Groups Viewing and changing Storage Group properties Adding and removing LUNs Connecting and disconnecting hosts Destroying Storage Groups LUN Masking with Access logix All LUNs are accessible through all SP ports LUN ownership is active/passive LUNS are assigned to storage Groups
When a host is connected to a storage group, it has access to all LUNs within the storage Group Access Logix Switch Zoning Zoning determines which hosts see what ports on a storage system Fabric level access control Multiple Hosts may be zoned to share the same ports Access Logix determines which LUNs are accessible to which host LUN level access control Both Zoning and Access Logix are used together
Access Logix Limits Host may be connected to only one Storage Group per array If multiple arrays in environment, host may be connected to one Storage Group in each array
Number of hosts per storage system varies based on the number of connections Maximum of 256 LUNs in a Storage Group A Storage Group is local to 1 storage system
Host agent must be running . If not, manually register initiators User must be authorized to manage Access Logix Power path What is Power path
Host Based Software Resides between application and SCSI device driver Provides Intelligent I/O path management Transparent to the application Automatic detection and recovery from host-to-array path failures The Value of Power path Support for Windows, and UNIX server Improves SAN performance Provides Path Failover
Allows applications to continue to access LUNS in the event of failure of any component in the IOP Requires careful planning and design to eliminate any single point of-failures Multiple HBAs Fabric zoning to provide multiple paths Provides load balancing Balances IO requests across HBAs and paths Does not balance IO across Storage Processors Supports EMC Symmetrix, CLARiiON and some 3rd party storage systems
PowerPath creates a path set for each LUN, and creates a pseudo-device that may be used in place of the native device EMC Power path A EMC Power path
A P S P A EMC POWER 0
P SCSI Device Driver LUN 0 S P B
EMC Power path Clariion Architecture CLARiiON supports an Active-Passive architecture When LUNs are bound, a default LUN owner is assigned
In the event of a SP or path failure, LUNs can be trespassed to the peer Storage Processor Trespass is temporary change in ownership
When the storage system is powered-on, LUN ownership returns to the Default Owner LUNs are owned by a Storage Processor Path Failover EMC power path Power Path Administration
Power path settings on Clariion for each host: Tools Failover setup wizard (Enable Array coman path and set Failover mode as 1 for power path. Power Path Administration provides both GUI(windows) and CLI (All platforms) CLI Administration: 1.Install Power path pkg on Hosts 2. Update PATH variable with /etc extension for all powerpath cmds to work 3.Add power path License: # /etc/emcpreg -add License Key # /etc/emcpreg list to list the installed power path license details 4. To verify that PowerPath devices are configured on the host:
# powermt display dev=all 5. To Configure any missing logical devices. #powermt config 6. To remove dead paths #powermt check 7. Powermt restore To restore dead paths after have been repaired Clariion Applications Clariion Applications
Snapview Snapshots Snapview Clones Mirror Copy Snap view over view
Snap view helps to create point-in-time copies of data Provide support for consistent on-line backup or data replication Data copies can be used for purposes other than backup (testing, decision support Scenarios) Snap view components: Snapshot Use pointer-based replication and Copy on First Write technology Make use of a Reserved LUN Pool to save data chunks Have three managed objects Snapshot, session, Reserved LUN Pool Clone
Make full copies of the source LUN Track changes to source LUN and clones in the Fracture Log Have three managed objects: Clone Group, Clone, Clone Private LUN Clones and Snapshots are managed by Navispheare Manager and Navisphere CLI Snapview Snapshots Snapshot Definition SnapView Snapshot - an instantaneous frozen virtual copy of a LUN on a storage system Instantaneous
Snapshots are created instantly no data is copied at creation time Frozen Snapshot will not change UNLESS the user writes to it Original view available by deactivating changed Snapshot
Virtual copy Not a real LUN - made up of pointers, original and saved blocks Uses a copy on first write (COFW) mechanism Requires a save area the reserved LUN Pool Snapview Snapshot Snapview Snapshot Components:
Reserver LUN pool Snapview Snapshot Snapview Session Production Host Backup Host Source LUN
Copy on First Write (COFW) Snapview Snapshot Components Reserved LUN pool: Collection of LUNs to support the pointer-based design of Snapview .Total number of reserved LUNs is limited.The limit is model-dependent. Snapview Snapshot: A defined virtual device that is presented to host and enables visibility into running session. The snapshot will be defined under a source LUN. Snapshot can only be assigned to single session.
Snapshot Session: Process of defining point-in-time designation by invoking copy-on-first-write activity for updates to the source LUN. Starting a session assigns a reserved LUN to the Source LUN. As far as this session is concerned, until a snapshot is activated ,the point in-time copy is not visible to any servers. At any time we can activate a snapshot to this session in order to present the point-in time image to a host. Each source LUN can have upto eight sessions Snapview Snapshot Components
Production Host: Server, where Customer Application execute Source LUNs are accessed from Production Host Backup Host: Host where Backup process occurs Backup Media attached to Backup Host Snapshots are accessed from Backup Host Source LUN: The LUN contains production data on which we want to start a Snap view Session and optionally activate a snapshot to that session COFW: The copy on first write mechanism involves saving an original data area from the source LUN into a Reserved LUN area when that data block in the active file
system is about to be changed Rollback: Enables recovery of the source LUN by copying data in the reserved LUN back to the Source LUN Managing Snapshots Procedure to Create and Manage Snapshots: 1. Configure Reserve LUN pool ReserveLUNpool- configure Add LUNs for both SPs 2. Create Storage group for prod host and add source LUN 3. Create file system on Source LUN and add data 4. Create Snapshot from LUN0
Storagegroup SourceLUN SnapviewCreate Snapshot 5. Create Snap session from LUN0 Storagegroup SourceLUN Snapview- Start SnapView session 6. Activate Snapshot Snapview- SnapshotsSelect the snapshot Activate Snapshot (Select a session for that snapshot) Managing Snapshots 7. Create Storage group for Backup host and add snapshot virtual LUN
8. Mount emc device of snap LUN on backup host 9. Verify the Data. 10 Do some modification from Prod Host 11. Umount the prod LUN
12. Perform Roll Back of Snap view session Snapview sessions Select sessionstart Rollback 13. Remount the prod LUN and observer the old data Snapview Clones SNAP view Clone SnapView Clones - Overview SnapView Clone a full copy of a LUN internal to a storage system. Clones take time to populate (synchronize)
Clone is independent of the Source once synchronization is complete 2-way synchronization Clones may be incrementally updated from the source LUN
source LUNs may be incrementally updated from a clone Clone must be EXACTLY the same size as source LUN Managing Snapview Clones Procedure to create Clones: 1. Prepare Clone private LUN(CPL) and Fractured log Storage system - Snapview - clone feature properties ( add those private LUNS) Fracture Log: Located in SP Memory Bitmap
Tracks modified extents between source LUN and each clone Allows incremental resynchronization in either direction Private LUN for each SP Must be 128 MB (250,000 blocks) or greater Used for all clones owned by SP No clone operations allowed until CPL created
Contains persistent Fracture Logs Managing Snapview Clones 2. Create Storage group for a host and add source LUN 3.Create file system for the emc device and add data 3. Create Clone group for a source LUN Storage System Snapview Create Clone group 4. Add clone to Clone group (Make sure the Synchronized status) SnapView ClonesClonegroup add clone 5. Fracture the Clone SnapView Clonesclone Fracture
6. Add clone to the Backup Host storage group. 7.Mount the emc device of the clone on Backup host and check the data. 8.Add some data on clone through backup host. 9. Initiate the Reverse Synchronization and observe the updated data from prod side Mirror Copy Mirror view Agenda Types of Mirror copy Synchronous ( Mirror view/S)
Asynchronous (Mirror view/A) How MirrorView make remote copies of LUNs The required steps in MirrorView administration Mirror View with Snap View Mirror Copy overview Optional storage system-based software
This product is designed as storage system-based disaster-recovery(DR) solutions for mirroring local production data to a remote/disaster recovery site. Mirrorview/S is a sysnchronous product that mirrors data between local and remote storage systems
Mirrorview/A is asynchronous product that offers extended-distance replication based on periodic incremental update model mirrors data Business requirements determine the structure of DR solution The buisiness will decide how much data loss is tolerable and how soon the data must be accessable again in the event of disaster.
Mirror copy overview It is a requirement that critical business information always be available. To protect this information it is necessary for a DR plan to be in place to safe guard against any disaster. Recovery objects: Recovery objects are service levels that must be met to minimize the loss of information and revenue in the event of disaster. The criticality of business application and information defines the recovery objectives. The terms commonly used to define the recovery objectives are:
Recovery point objective(RPO) Recovery time objective(RTO) Recovery point objective: Recovery point objective defines the amount of acceptable data loss in the event of disaster. RPO is typically expressed in duration of time. Some applications may have zero tolerance for loss of data in the event of disaster. (Example: Financial Applications)
Mirror copy overview Recovery time objective(RTO): RTO is defined as amount of time required to bring the business application back online after disaster occurs. Mission critical application may be required to be back online in seconds, without any noticeable impact to the end users. Replication Models Replication solutions can be broadly categorized as synchronous and asynchronous. Synchronous replication model: In a synchronous replication model, each server write on the primary side is written
concurrently to the secondary site. RPO is zero, since the transfer of each I/O to the secondary occurs before acknowledgement is sent to the server Data at the secondary site is exactly the same as data at the primary site at the time of disaster disaster. Asynchronous replication model Asynchronous replication models decouple the remote replication of the I/O from the acknowledgement to the server. Allows longer distance replication because application write response time is not
dependent on the latency of the link. Periodic updates happens from primary to secondary at user-determined frequency Biderection Mirroring Mirror view Terminology and Data States Primary Image: The LUN contains production data and contents of which replicated to the secondary Image Secondary Image: A LUN that contains the mirror of the primary image LUN.This LUN must reside on different clariion than the primary Image. Fracture: A condition in which the I/O is not mirrored to the secondary image. (Admin Facture,System Facture)
Promote: The operation by which the administrator changes an images role from secondary to primary. As part of this operation the previous primary will become secondary. MirrorView/S Fracture Log and Write Intent Log Fracture Log:
Resident in SP memory, hence volatile Tracks changed regions on Primary LUN when Secondary is unreachable When Secondary becomes reachable, Fracture Log is used to resynchronize data incrementally Fracture Log is not persistent if Write Intent Log is not used Write Intent Log:
Optional allocated per mirror Primary LUN Persistently stored - uses private LUNs Used to minimize recovery in the event of failure on Primary storage system Two LUNs of at least 128 MB each MirrorView Mirror Creation Connect storage systems Physically, by zoning Logically, by Manage MirrorView Connections dialog Create Remote Mirror
Designate a LUN to be a Primary LUN Specify a mirror name and a mirror type Add secondary image(s)
Mirror is created in the inactive state, quickly changes to active Remote mirror view connection Configure and Manage Mirror view/S 1.Add source LUN to storage group and create file system on it and store some data. 2.Manage mirror connections. Storage system -- mirrorView--- >Manage mirror connections 3. Allocate write intent log Storage system MirrorView- Allocate write intent log 4.Create Remote mirror
Storage System-- MirrorView- Create Remote Mirror 5. Add Secondary Image Remote Mirrors Select the mirror Add Secondary Image 6. Promote the secondary and add the LUN to the any DR storage group and verify the data.
n order to familiarize new markets with its bicycles, Raleigh brings samples to cities around the world and lets people ride the bikes. Raleigh takes great strides to remain in touch with its target market and to provide the products...
Intentionally making partner feel jealous is positively correlated (.15 - .33) with relational improvement (Fleischman et al., 2004) 85% reported their relationship stayed the same or improved as a result of the last episode of violence (Gryl et al., 1991)
Why I Like Spiders Spiders Eat Annoying Bugs Spiders Build Cool Webs Spiders Are Not Insects Most Spiders Are Harmless Tarantulas Make Good Pets Spiders Eat Annoying Bugs Mosquitoes House flies Cockroaches Gnats Small Household Pets (Just Kidding) Spiders Build...
List of 2017 campus comparison groups released (TEASE and public web) June 15. Confidential lists of college and career ready graduates for 2017 state accountability released (TEASE) 2017 Accountability Calendar. Date. Activity; Monday, August. 7.
players and to signal the expiration of the period of exclusion by raising the appropriate flag. PROPOSED rule: The use of automatic visual effects, surrounding the penalty box. i.e. signalling. a) the last 5 seconds of the expiration of the...
Speaker's Notes: The point of operation, or the area where the work takes place, is the primary hazard of any power tool. The powered and moving saw blade, grinding wheel, wire brush, sander, drill bit, chisel, etc., will cause a...
Primera descripción del algoritmo fue dada por Werbos en 1974 Generalización del algoritmo de Widrow-Hoff para redes multicapa con funciones de transferencia no-lineales y diferenciables. 1989 Hornik, Stinchcombe y White Una red neuronal con una capa de sigmoides es capaz...
Ready to download the document? Go ahead and hit continue!