What is “Tier 0” in Storage Environments?

Tier "0" is not new in storage market but for implementation purposes it has been difficult to accommodate because it requires best performance and lowest latency. Enterprise Flash disks (Solid State Disks) capable to meet this requirement. It is possible to get more performance for company most critical applications. The performance can be gained through using Flash drives supported in VMAX and DMX-4 systems.  Read More →

EMC introduced new feature for CLARiiON with Flare release 29

Posted by Diwakar ADD COMMENTS

CLARiiON Flare release 29 (04.29.000.5.001) introduce support for several new features as follow:

1) Virtualization-aware Navisphere Manager - Discovery of VMware client always were difficult in earlier release but Flare 29 enables CLARiiON CX4 users and VMware administrators to reduce infrastructure reporting time from hours to minutes. Earlier releases have allowed only a single IP address to be assigned to each iSCSI physical port. With Flare 29, the ability to define multiple virtual iSCSI ports on each physical port has been added along with the ability to tag each virtual port with unique VLAN tag. VLAN tagging has also been added to the single Management Port interface. It should be noted that the IP address and VLAN tag assignments should be carefully coordinated with those supporting the network infrastructure where the storage system will operate

2) Built-in policy-based spin down of idle SATA II drives for CLARiiON CX4 - Lowers power requirements in environments such as test and development in physical and virtual environments. Features include a simple management via a “set it and forget it” policy, complete spin down of inactive drives during times of zero I/O activity, and drives automatically spin back up after a "first I/O" request is received.

3) Virtual Provisioning Phase 2 - Support for MirrorView and SAN Copy replication on thin LUNs has been added.

4) Search feature – Provides users with the ability to search for a wide-variety of objects across their storage systems. Objects can be either logical (e.g., LUN) or physical (e.g., disks).

5) Replication roles - Three additional roles have to be added in Navisphere: “Local Replication Only”, “Replication” and “Replication/ Recovery”.

6) Dedicated VMware software files - VMware software files (i.e. NaviSecCLI, Navisphere Initialization Wizard) are now separate from those of the Linux Operating System.

7) Software filename standardization - all CLARiiON software filenames beginning with FLARE Release 29

8) Changing SP IP addresses - SP IP addresses can now be changed without rebooting the SP. Only the Management Sever will need to be rebooted from the Setup page, which results in no storage system downtime.

9) Linux 64-bit server software – Native 64-bit Linux server software files simplify installation by eliminating the need to gather and load 32-bit DLLs.

10) Solaris x64 Navisphere Host Agent – Release 29 marks the introduction of Solaris 64-bit Navisphere Host Agent software. This Host Agent is backward compatible with older FLARE release.

Read More
, , , , , ,

Table Of Post Content - On Subscriber Request.

Posted by Diwakar ADD COMMENTS


Read More
, , , , , ,

EMC Open Replicator:- Another option for Migration

Posted by Diwakar ADD COMMENTS

Lets understand EMC Open Replicator product:- Open Replicator enables remote point-in-time copies with full or incremental copy capabilities to be used for high-speed data mobility, remote vaulting, migration, and distribution between EMC Symmetrix DMX and other qualified storage arrays. OR leverages the high-end Symmetrix storage architecture and offers unmatched deployment flexibility and massive scalability.

EMC Open Replicator is tightly integrated with EMC TimeFinder and SRDF families of local and remote solutions. Open Replicator Functionality:

- Protect lower-tier applications at remote locations.

- Pushing or Pulling data from Symmetrix DMX arrays to other qualified storage arrays in SAN/WAN environments.

- Create remote point-in-time copies of local production volumes for many purposes from data vaulting to remote testing and development.

- Ensure immediate data availability to host applications via Symmetrix DMX consistency technologies.

When to use EMC Open Replicator for migration:


1) SYMAPI cannot validate third party or non-visible storage systems.
2) To protect against potential data loss due to a SAN failure or other connectivity issue during a hot pull operation, use the donor update option. When enabled, this feature causes all writes to the control device from the host to be immediately copied to the remote device as well. Because the data is fully copied to both the remote device and the control devices, if a failure occurs, the session can safely be terminated and created again to fully recover from any mid-copy failure.
3) Open Replicator uses FA resources. If you are using this utility in a production environment, verify with the SA that FA bandwidth assessments have been considered and that appropriate throttling parameters (pace or ceiling).
4) When using BCVs, the BCVs must be “visible” to the remote storage array. Thus, they must be mapped to an FA and the FA must be zoned to the destination storage. We highly recommend that BCVs not be mapped to the same FA as the control standard devices to avoid a negative impact on host I/O performance.
5) If a configuration uses thin device as destination in a pull or push copy operation, full volume allocation of the thin device will be made because Open Replicator creates a full volume copy.
6) When performing an Open Replicator migration, always use the –v qualifier on the create command. This insures that, should the session fail, there will be useful information returned as to what volume caused the error. This allows you to more quickly recognize zoning or masking issues.
7) Issuing create commands prior to Open Replicator migration activities allows confirmation that there will be no zoning or masking issues discovered during the migration window. This technique will only be successful if no changes have been made to the Symmetrix environment between issuance of the create and copy commands.
8) It is better and easier to use an Open Replicator management host for preparing, executing, and monitoring migration sessions than using one of the systems with volumes involved in the activity.
9) For Veritas file systems, PowerPath devices, and Oracle databases, when you activate the copy session devices must be frozen just before the activate is performed, and thawed as soon as the activate completes. Use the following options to with the symrcopy activate command, when applicable:
-vxfs MountPoint
-ppath srcdevs

PowerPath device

-rdb dbtype DbType -db DbName

10) The device specified in the command line must match the device in the device file or the activate will fail.

Read More
, , ,

PowerPath Migrator Enabler Configuration Consideration

Posted by Diwakar ADD COMMENTS

PowerPath Migrator Enabler is a host-base migration tools from EMC that allows you to migrate data between storage systems with little or no interruption to data access. This tool can be use in conjunction with other underlying technology such as EMC Invista, Open Replicator. PPME use the PowerPath filter dirvers to provide non-disruptive or minimally disruptive migrations. Only specific host plateforms are supported by PPME. Please check EMC support matrix for supported host systems. One of the PPME features that supports pseduo-to-pseduo, native-to-native and native-to-pseudo device migration.

Consider the following when designing and configuring PPME:

Ø Remote devices do not have to be the same RAID type or meta-configuration.
Ø Target devices must be the same size or large than the source control device.
Ø Target directors act as initiators in the SAN.
Ø Contrary to the recommendations for Open Replicator, the source device remains online during the “hot pull.”
Ø The two storage systems involved in the migration must be connected directly or through a switch, and they must be able to communicate.
Ø Every port on the target array that allows access to the target device must also have access to the source device through at least one port on the source array. This can be counter to some established zoning policies.
Ø Since PPME with Open Replicator uses FA resources, determine whether this utility will be used in a production environment. In addition, consider FA bandwidth assessments so that appropriate throttling parameters (that is, pace or ceiling).
Ø The powermig throttle parameter sets the pace of an individual migration by using the pace parameter of Open Replicator:
Ø Faster (lower throttle) makes the migration faster, but may impact application I/O performance.
Ø Slower (higher throttle) makes the migration slower.
Ø The default is five (midpoint).
Ø When setting a ceiling to limit for Open Replicator throughput for a director/port:
Ø The ceiling value is set as a percentage of a director/port’s total capacity.
Ø The ceiling can be set for a given director, port, director and port, or all director and ports in the Symmetrix array.
Ø To set ceiling values, you must use symrcopy set ceiling directly (powermig does not provide a way to do this)
Ø Once the hot pull has completed, remove or re-use the source device.
Ø Do not forget to “clean up” the zoning once you have completed migration activities.
Hope this will be useful in migration planning or selecting migration tools. I will try to explain in deatil in coming post such as PPME with Open Replicator, Solution Enabler etc..

Read More
, , , , , ,

What is “Tier 0” in Storage Environments?

Posted by Diwakar ADD COMMENTS

In the current storage markets and technology, storage tiers are defined by availability, functionality, performance and costs. In fact data can move up and down tiers as time and business required.

Tier "0" is not new in storage market but for implementation purposes it has been difficult to accommodate because it requires best performance and lowest latency. Enterprise Flash disks (Solid State Disks) capable to meet this requirement. It is possible to get more performance for company most critical applications. The performance can be gained through using Flash drives supported in VMAX and DMX-4 systems. One Flash drive can deliver IOPS equivalent to 30 15K RPM hard disk drives with approximately 1 ms application response time. Flash memory achieves performance and the lowest latency ever available in the enterprise class storage array.

Tier “0” application can be closely coupled with other storage tier within Symetrix series for consistency and efficiency, reducing cost of company for manual data layout or data migration from old disk to new high speed disk.

Tier “0” storage can be used to accelerate online transaction processing, accelerating performance with large indices and frequently accessed database tables i.e. Oracle, DB2 databases and SAP R/3. Tier 0 can also improve performance in batch processing and shorten batch processing in windows environments.

Tier “0” storage performance will help application that needs the lowest latency and response time possible. The following applications can get benefited through using Tier 0 storage:

- Algorithmic trading
- Data modeling
- Trade optimization
- Realtime data/feed processing
- Contextual web advertising
- Other realtime transaction systems
- Currency exchange and arbitrage

Tier “0” storage is most beneficial with random read miss application. If random read miss percentage is low, application will not see any performance difference since writes and sequential reads/writes already leverage Symmetrix cache to achieve the lowest possible response time.

For example, if the read hit percentage is high >90 % as compared to read misses, such application like DSS, Streaming media, improvements provided by Tier 0 storage will not likely be enough to be cost-effective.

Read More
, , , , ,

TimeFinder/Mirror Consistent Split

Posted by Diwakar ADD COMMENTS

Think when you are creating a point-in-time image for multiple devices. It is easy to create a point-in-time image of entire set of logical device at same time. In order achieve this you need to shut down an application so that no IO will occurs while you creating a point-in-time image. This is big problem in today’s environment where every company looking solution for zero down time.
The EMC provided solution to this problem is called “Enginuity Consistency Assist”. When you create a set of sessions and invoke Enginuity Consistency Assist, the Symmetrix aligns the I/O of those devices and halts all I/O from the host systems very briefly—much faster than the applications can detect—while it creates the session. It then resumes normal operation without any application impact.
TimeFinder Consistent Split using (TimeFinder/Consistency Groups) allows the splitting off of a consistent, re-startable image of an Oracle database instance within seconds with no interruption to the online Oracle database instance.
Ÿ - Allows users to split off a dependent write consistent, re-startable image of application without interrupting online services
Ÿ - Using TimeFinder/Consistency Groups to defer write I/O at the Symmetrix before a split
Ÿ - Consistent split can be performed by any host running Solutions Enabler connected to the Symmetrix
Ÿ - Tested and available including HP-UX, Solaris, AIX, Linux, and Windows
Ÿ - No database shutdown or requirement to have database put into backup mode (Oracle).

Using TF/CG, consistent splits helps to avoid inconsistencies and restart problems that can occur with using Oracle hot-backup mode (not quiescing the database).
The major benefits of TF/CG are:
• No disruption to the online Oracle database to obtain a Point-in-Time image
• Provides a consistent, re-startable image of the Oracle database for testing new versions or database patch updates before deploying for use in production environments
• Can be used to obtain a business point of consistency for business restart requirements for which Oracle has been identified as one of multiple databases for such an environment.

The same benefits apply using TF/CG in a clustered environment as in a non-clustered environment:
- No disruption to the online Oracle database to obtain a Point-in-Time image in a Oracle single instance environment or when using Oracle Real Application Clusters
- Provides a consistent, re-startable image of the Oracle database for testing new versions or database patch updates before deploying for use in clustered production environments
- Can be used to obtain a business point of consistency for business restart requirements for which Oracle has been identified as one of multiple databases for such an environment.

Read More
, , , , ,

Understanding Auto Provisioning in VMAX

Posted by Diwakar ADD COMMENTS

Auto-provisioning requires Enginuity 5874 or later. It simplify Symmetrix provisioning by allowing you to create group of devices like storage group in CLARiiON, Front-End Port Group and Host Initiators Group and then associate these groups with each other in a masking view.

The following are the basic steps for provisioning Symmetrix using Auto-Provisioning:-

1) Create a Storage Group
2) Create a Port Group
3) Create an Initiator Group
4) Associate the groups in a Masking View.

Creating Storage Group:- It is component of Auto-Provisioning group and FAST ( Will discuss about FAST in later post), both require Enginiuity 5874. The maximum number of storage group allowed per array is 8192. A storage group can contain up to 4096 devices. A Symmetrix device can belong to more than one storage group.

Note:- By default Dynamic LUN addresses will assigned to each device. If can manually assign the host LUN addresses for the device you are adding to the group by clicking Set LUN Address- Storage group dialog box.

Creating Port Group:- A port can belong to more than one port group and port must have the ACLX bit enabled. For example if you want FA 5A and 12 A for windows operating system, you can create port group name called WIN_PortGrp or Win_FA5A_FA12A_PrtGrp etc.

Creating Initiator Group:- The maximum number of initiator groups allowed per Symmetrix array is 8000. An initiator group can contain up to 32 initiator of any type and contain other initiator groups (cascaded to only one level).

Initiator Group name must be unique from other initiator groups on the array and cannot exceed 64 characters. Initiator group names are case-insensitive.

Creating Masking view:- It just a co-relation between Storage Group, Port Group and Initiator Group and you are done! Device will be mapped automatically to selected port group and masked to selected initiator groups.

Read More
, ,

About Me

My photo
Sr. Solutions Architect; Expertise: - Cloud Design & Architect - Data Center Consolidation - DC/Storage Virtualization - Technology Refresh - Data Migration - SAN Refresh - Data Center Architecture More info:- diwakar@emcstorageinfo.com
Blog Disclaimer: “The opinions expressed here are my personal opinions. Content published here is not read or approved in advance by EMC and does not necessarily reflect the views and opinions of EMC.”
EMC Storage Product Knowledge Sharing