Archive for December, 2013

What to consider when planning a cluster relocation

Kirsten Gantenbein

Kirsten Gantenbein

Principal Content Strategist at EMC Isilon Storage Division
Kirsten Gantenbein
Kirsten Gantenbein

Moving an EMC Isilon cluster to a new location, whether it’s in your current building or several miles away, can be complicated. If you’re considering a physical move of your cluster, contact your Isilon account representative to have EMC Professional Services help you with the entire process, from planning to reinstalling your cluster at the new location.

This article highlights three important factors to consider before planning your move.

TimeframeMoving

An EMC Project Manager will be assigned to manage your cluster relocation project. Once the project is underway, it’s recommended that you allow at least a four-week window in which to plan and perform your move. Four weeks before the move, you should work with your Project Manager to confirm moving dates, document the serial numbers of all nodes and InfiniBand switches, and determine the shipping packing materials that you might need.

If you want to make configuration changes to the cluster or use new IP addresses at the new location, start planning for how you want the cluster to be reconfigured. An EMC Solutions Architect can assist with this discussion about planning a cluster relocation, and gather the necessary information. Two weeks before the move, you should perform a cluster maintenance check and back up your data. After the move is complete, EMC engineers to reinstall and validate your cluster you should factor in time for.

Site preparation

To help plan for the cluster relocation, read the Isilon Site Preparation and Planning Guide, available through the EMC Online Support site. This guide will help you prepare the destination site for the cluster, and provides guidelines about the following site requirements:

  • Power
  • Square footage
  • Level floors
  • Temperature and humidity
  • Network capacity
  • Cabling
  • IP addresses and cluster configuration
  • Rack space

If you have the original Isilon shipping boxes, they can be used for the cluster relocation. Otherwise, two weeks before the move, your account representative will order Isilon-approved shipping boxes for the move. At this time, you should confirm that the new location will be ready to accept the cluster shipment and has the space and infrastructure necessary for unpacking and reinstalling your cluster.

Data backup and cluster health validation

It is critical to back up your data to a secure location before the move. It is also important to validate your cluster’s health to address any issues before the move. You can open a Service Request (SR) for EMC Isilon Technical Support to review your logs and validate your cluster’s health before the move. Alternatively, a Solutions Architect or Project Manager can assist you with opening the SR.

Both of these steps should be done during a two-week maintenance window preceding the move.

For more information

Professional Services has the expertise to ensure that your cluster move goes as smoothly as possible. If you have questions about the EMC Isilon Cluster Relocation service, please contact your Isilon account representative.

Data Protection and Disaster Recovery with EMC Isilon SyncIQ

Kirsten Gantenbein

Kirsten Gantenbein

Principal Content Strategist at EMC Isilon Storage Division
Kirsten Gantenbein
Kirsten Gantenbein

To prepare for disaster scenarios, such as an unplanned outage that can potentially compromise your data, you need a solution that creates and safely stores copies of your data.

EMC Isilon SyncIQ is Isilon’s data replication software and a key element of a robust data protection and disaster recovery solution. SyncIQ works with the underlying snapshot technology of EMC Isilon SnapshotIQ software to copy data blocks that change between replications. The replication process is completed quickly and with minimal disruption.

In this video, Principal Solutions Architect Amol Choukekar, answers frequently asked questions about SyncIQ. Watch this video to gain a basic understanding of the technical concepts and functionality of SyncIQ. For more information, contact your EMC Isilon account representative.

Video Transcript

Hello. This is Amol Choukekar, Principal Solutions Architect with EMC Isilon storage division.

Today, I’m going to talk about EMC Isilon SyncIQ and how it helps in implementing a disaster recovery and data protection strategy.

In this video, we’ll talk about EMC Isilon SyncIQ’s design features, basic functionality, and the operations associated with it.

SyncIQ is EMC Isilon’s data replication feature.

It is a licensed feature of OneFS and it replicates data from one cluster to another cluster, asynchronously, over a standard LAN or WAN connection.

SyncIQ allows creating user-defined policies that can be scheduled to replicate data to meet data recovery point objectives.

SyncIQ maintains the integrity of data by copying the data as well as the metadata, like permissions associated with files and directories.

SyncIQ uses underlying snapshot technology that performs block-based delta copies between replications.

This results in tremendous WAN bandwidth savings and shrinks the time needed to keep the data in sync between two clusters.

SyncIQ is typically set up to replicate data from a single source cluster to a single target cluster in a one-to-one relationship.

However, SyncIQ is capable of replicating from a single source cluster to multiple target clusters in a one-to-many relationship.

As we can see in this example here, we have a dataset, D1, replicating to both cluster 2 and cluster 3 in a one-to-many relationship.

At the same time, we can see that we have cluster 1 and cluster 3 with independent datasets participating in a SyncIQ relationship in a bi-directional fashion.

Thus, we can say that SyncIQ can perform one-to-one, one-to-many, and bi-directional replication with independent datasets.

Starting with EMC Isilon OneFS 7.0, SyncIQ functionality has been enhanced for ease of failover and failback.

Now, let’s take a look at the operations associated with SyncIQ.

As we can see in the example on the whiteboard here, we have a production side and we have a disaster recovery side. We have data on the production side in read-write mode and we have data on the disaster recovery side as read-only. This is marked read-only to maintain the integrity of the data.

Now, let’s talk about failover. Failover can be temporary or failover can be permanent in case of a disaster.

Now, let’s talk about a temporary failover. For example, a DR testing scenario.

What is failover? Failover is nothing but marking the target directory on the disaster recovery side as read-write.

Now, let’s say in case this was temporary, and we are done with the disaster recovery test, we wrote some data, D2, as in this example, and we want to continue the normal operation as it was before.

What would need to happen is a process called as a “failover revert”.

What is failover revert? Failover revert is nothing but discarding the changes that were written while we were in a read-write state and marking the target side back as read-only so that normal processing can continue.

Now, what happens in case of an actual disaster?

You would have your production side inaccessible.

You’ll initiate a failover process which marks the target side as read-write and you would point your clients to the target side or the disaster recovery side, and you continue writing new data to it, D2, D3, in this case.

Let’s say after a period of time, the production side comes back up and we want to go back to the original production, there’s a couple of processes that need to happen.

While we were on the disaster recovery side, we wrote some new data, which is significant data now.

That needs to get copied back to the production side and that is done through what is called as a “failback” process.

So, what is a failback process? Failback process is reversing the direction of the SyncIQ policy and copying all the changes from the disaster recovery side to the production side.

So you would have your D2 and your D3 copied to the production side.

Once that is done, a normal failover can be initiated, where we can mark the production side as read-write and we can continue your normal operations in the same direction from production to disaster recovery side and mark disaster recovery side back as read-only and continue replicating data to it.

As we saw in this video, SyncIQ provides powerful functionality for disaster recovery and data protection.

If you have any further questions about SyncIQ and would like to implement SyncIQ in your environment, please feel free to contact your EMC account team or EMC Business Partner.

Thank you for watching.

The top 20 support documents in November 2013

Kirsten Gantenbein

Kirsten Gantenbein

Principal Content Strategist at EMC Isilon Storage Division
Kirsten Gantenbein
Kirsten Gantenbein

One of the goals of this blog is to share the most useful EMC Isilon support-related content that we have to offer. In this post, we’re highlighting 20 of the most viewed knowledgebase (KB) articles and product documents from the month of November.

We hope these documents will help you to quickly find an answer to a common question or resolve an issue.

Top 10 KB articles

To access these KB articles, log in to EMC Online Support.

  1. Best practices for NFS client settings (90041)
  2. OneFS pre-upgrade preparation (88672)
  3. Troubleshooting performance issues (88844)
  4. Matrix of patches available for Isilon OneFS (88538)
  5. How to reset a node to factory defaults (16696)
  6. How to upload files to Isilon Technical Support (16759)
  7. OneFS sysctl commands (89334)
  8. How to create a bootable image of OneFS on a USB flash drive (16691)
  9. Visio Stencils of Isilon Clustered Storage Systems (90170)
  10. How to reimage a node using a USB flash drive (16582)

 

Top 10 product documents

To access these PDF documents, log in to EMC Online Support.

  1. Current Isilon Software Releases
  2. Isilon Supportability and Compatibility Guide
  3. OneFS 7.0.1 Administration Guide
  4. OneFS 7.0.2 Administration Guide
  5. InsightIQ 2.5 Installation and Setup Guide
  6. Isilon Node Firmware Package 8.3 Release Notes
  7. OneFS 7.0.2 Command Reference
  8. Insight 2.5 User Guide
  9. OneFS 7.0.1.10 Release Notes
  10. OneFS 6.5.5 Command Reference

 

A closer look at EMC Isilon SmartDedupe and the Isilon OneFS 7.1 Job Engine

Kirsten Gantenbein

Kirsten Gantenbein

Principal Content Strategist at EMC Isilon Storage Division
Kirsten Gantenbein
Kirsten Gantenbein

We recently learned that our blog readers are most interested in the new EMC Isilon SmartDedupe software offering and the Job Engine enhancements incorporated into in the recently released EMC Isilon OneFS 7.1.

In this video, we take a closer look at these features. First, we cover basic concepts about SmartDedupe and Job Engine performance enhancements in OneFS 7.1. Next, we provide brief demonstrations of how to use these features in the OneFS web administration interface. This video also highlights details about data at rest encryption in OneFS 7.1.

Download the EMC Isilon OneFS 7.1 release notes for more information. You can also review the video transcript below.

Video Transcript

Hello, I’m André Morrissen, a Senior Technical Writer with the Information Development team.

Version 7.1 of OneFS contains numerous enhancements that will improve the performance of your Isilon cluster.

In this video, we’ll take a look at a SmartDedupe, job engine improvements, and data at rest encryption and find out how they can improve your workflow.

SmartDedupe is a new licensed feature of OneFS which enables you to save storage space on your cluster by reducing redundant data—in other words, by deduplicating that data.

SmartDedupe is most beneficial for workflows that incorporate large amounts of duplicate data, such as archiving or when a large amount of virtual machines are stored on a cluster.

As you write files to the cluster, some of those files or blocks of data in the files might be duplicates. You can run a deduplication job that scans the file system to see if that data already exists. If it does, OneFS moves that data to a hidden file called a shadow store and replaces the duplicate data in the files with a pointer to the shadow store.

Deduplication is applied at the subdirectory level and targets all files and directories underneath one or more root directories.

The deduplication job is set to run at low priority by default, so impact to your workflow should be minimal. However, it’s a good idea to wait until users have finished modifying their files on the cluster before you run the job.

You can perform the following deduplication tasks from the OneFS web administration interface.

Assign specific subdirectories for deduplication.

Run an assessment job to determine how much space you might save in a given directory.

And view detailed reports of deduplication jobs.

OneFS 7.1 includes major improvements to the job engine, the system that helps you schedule and manage maintenance jobs on your cluster.

As with previous versions of OneFS, the job engine can adjust jobs based on the amount of cluster resources available. For example, if clients require more system resources, threads allocated to the job engine are decreased.

However, now you can run up to three jobs simultaneously, with a few exceptions that keep similar types of jobs from colliding. For example, you can run an AutoBalance, IntegrityScan, and DedupeAssessment job all at the same time.

OneFS 7.1 also introduces support for Data at Rest Encryption.  With this feature, you’ll be able to create a cluster of nodes that contain self-encrypting drives or SEDs. Data at Rest Encryption provides data security that meets specific regulatory requirements for financial and governmental workflows.

Isilon’s use of hardware-based encryption provides the following benefits:

Less consumption of system resources.

Removed drives remain encrypted, which prevents data theft.

The data encryption is performed at the drive level using special processors on each SED that provide 256-bit AES encryption protection. The encryption has less than 1% impact on the performance of the drives themselves.

If you’re interested in creating a cluster with SEDs and using Data at Rest Encryption, contact your account representative.

For more information about the features in this videos, see the OneFS Web Administration.

For a full list of new features, see the OneFS 7.1 Release Notes.

Thanks for watching.

How to start a job in the EMC Isilon OneFS Job Engine

Kirsten Gantenbein

Kirsten Gantenbein

Principal Content Strategist at EMC Isilon Storage Division
Kirsten Gantenbein
Kirsten Gantenbein

Optimizing free space on your Isilon cluster.

Scanning the file system to ensure that data is protected.

Removing duplicate data.

These are examples of cluster maintenance tasks that are managed by the Job Engine, a work distribution service in the EMC Isilon OneFS operating system.

The Job Engine plays an important role in maintaining the reliability and integrity of your cluster, and is well documented on the EMC Online Support site. It’s also a popular feature with customers, as indicated by our recent poll results.

This post provides a brief overview of the Job Engine and a video demonstration of how to start a job.

Job Engine basics

The Job Engine prioritizes, schedules, and balances the amount of resources dedicated to a specific set of tasks, or jobs, that run in the background. Specifically, the Job Engine distributes job segments—broken into phases and tasks—across the nodes in your Isilon cluster. For a list of cluster maintenance jobs that are managed by the Job Engine, see the OneFS administration guides or the knowledgebase article titled OneFS 5.0 – 7.0: Complete list of jobs by OneFS version. (Log in to the EMC Online Support site to view this article and OneFS administration guides.)

You can review the schedule of jobs and monitor the progress of running jobs in the OneFS web administration interface or command-line interface. Because some jobs can take many hours to complete or consume a significant amount of resources while running, you can also manually start, pause, or stop jobs. If a job is paused, you can restart the job at any time from the point at which it was interrupted.

In OneFS versions 6.0.x through 7.0.x, only one job can run at a time in the background. In OneFS 7.1, up to three jobs can run at the same time (with a few exceptions to keep similar types of jobs from interfering with each other). Each job has a default impact policy and priority to determine when it will run. You can also modify these default settings when you start a job.

How to start a job

If you want to start a job before it is scheduled to run automatically, you can start it manually. This basic procedure for OneFS 6.0.x through 7.0.x is described in the following video.

To start a job in OneFS 7.1, go to Cluster Management > Job Operations > Job Types click More (circled in red below), and then click Start Job.

How to start a job in Isilon OneFS 7.1

How to start a job in OneFS 7.1.

Additional details about the Job Engine in OneFS 7.1 can be found in the recently updated white paper titled EMC Isilon OneFS: A Technical Overview.

Are you looking for more specific information about job management in OneFS? Let us know. Send an email to isi.knowledge@emc.com, or post a comment on this blog.