THE ZETTA.NET BLOG

Cloud Backup & Disaster Recovery
News and Info

Posts Tagged ‘data loss’

Courtenay Troxel

June 13, 2012

Affordable Disaster Recovery: Despite Complexity, Tapes Work

Courtenay is a Channel Marketing Strategist at Zetta.net.

Disaster Recovery Lemons
Preparing for a data disaster is one of those projects that are always getting pushed back in small and medium sized businesses. The daily proverbial “fires” seem to take precedence over getting ready for an actual fire, server crash, earthquake, theft, or hurricane.
 
Implementing a data disaster recovery strategy is critical, but is often delayed for 2 reasons:
 
•   The complexity of evaluating business operations to find the critical data that needs to be made available first after a disaster, then calculating the target RPO and RTO for that data.
 
•   The perception that a disaster recovery solution that meets targets is too expensive
 
As Storage Switzerland noted recently, tapes are holding their position as an affordable backup solution that can also be used for disaster recovery. So, if you are using tape for backup here’s how to set up for recovery after a data disaster:
 

1. Optimize Tape Rotation For Recovery

 
Consider keeping Monday, Tuesday, Wednesday and Thursday tapes with full overnight backups of everything and taking them off site each day. These are the tapes you’d restore in a disaster recovery scenario.
 
Then, depending on the needs of your business, have another set, for example each Friday of the month. They can stay onsite in a fireproof safe, for recovering an accidentally deleted folder, for example. Most safes aren’t melt proof, so after a certain amount of time the tapes will still be vulnerable, so remember to consider the level of heat protection being offered when you’re shopping for a safe.
 
Then, for archival purposes, make 6 (or 3 or 9) monthly tapes, that can stay in the same fireproof safe. Finally, once a year, run a backup and take it to a bank safety deposit box.
 

2. Consider The Recovery Environment

 
This is the part that’s hard to generalize about. Depending on your geography and environment, think through the factors that will limit your recovery capabilities after a disaster.
 
Are you in a tornado prone area? It may take days to receive a new server after a storm damages road and communication infrastructure in the area. In this case, an extra auto-loader and server could be kept offsite with the daily backup tapes. The offsite storage location might also be where you’d personally go during an evacuation, for example.
 
In a fire or theft scenario, new hardware can be overnighted, but that still puts company data at a 24-36 hour recovery window. This may be a manageable risk or totally unacceptable, depending on the nature of the business.
 
As the title of this post suggests, using tape backups for disaster recovery can be pretty complex, but are a good way to leverage existing equipment to get a DR strategy in place.
 
Of course, with Zetta’s DR-as-a-Service protecting data offsite can be done over the web in a matter of minutes, with no-touch daily backups and instant recovery. Tapes will work fine, it’s just that Zetta works better.

Courtenay Troxel

May 17, 2012

When Pixar Deleted Toy Story 2 – How Cloud Disaster Recovery Makes a Hero of IT

Courtenay is a Channel Marketing Strategist at Zetta.net.


 

The video above tells the story of the day that someone at Pixar accidentally ran the “RM*” delete command on the Linux and Unix workstations that held all the Toy Story 2 data. The work of 30 people over a year was deleted in :20 seconds. As tends to happen in data disasters, the local backup system had also been failing for a month without being noticed, so the film was gone.
 
In the end, a local copy of the data was found on the technical director’s personal laptop at home, and the movie went on to earn $500 million at the box office and a rare 100% on the Tomatometer.
 

What can we learn from Pixar’s scenario that can benefit other mid-sized enterprises with large amounts of critical data?

 
1. The Importance of online /cloud backups
 
The list of reasons not to rely on local backups is long. It includes physical damage like floods and tornadoes that destroy primary data sources and backups at the same time, as well as human errors like in this Pixar case.
 
2. Use replicate and sync to ensure that the whole file system is replicated in the cloud
 
When critical data, like graphics and video effects files, are accidentally deleted, they need to be recovered quickly so people can get back to work. The best way to do this is with a replicated file system in the cloud. Having backup data in a proprietary backup format doesn’t help when it’s time to recover that data for use.
 
Additionally, any solution that doesn’t let you instantly check the validity of your backups should be considered a deal-breaker. Basically, this means all tape backup systems.
 
3. Recover quickly via the Internet to dissimilar hardware, even if the datacenter is down
 
When you’re dealing with a live data disaster recovery scenario, as opposed to a recovering a single file or folder, having web access to replicated backup files that remote employees can work from means business can continue.
 
4. Importance of multi-platform support
 
The era of the all “Windows shop” is quickly ending. A backup solution simply must support Linux and Mac OS in addition to Windows. Zetta for example, supports 18 platforms, including 11 different versions of Linux, 5 versions of Windows, and 2 versions of Mac OS.
 
5. Importance of supporting remote users and offices
 
VPNs and ad-hoc FTP servers have been a few of the somewhat reliable and secure options for backing up data from remote locations in the past. The evolution of enterprise-grade cloud solutions, like Zetta, now allows you to backup data from remote offices and users in minutes, without the hassle of VPN connections, appliances or servers.
 
The importance of having a recovery optimized backup can’t be understated. When three little characters like “RM*” can erase $500M worth of data, that little monthly fee for cloud backup becomes worth it for the peace of mind alone.

Courtenay Troxel

April 03, 2012

Tape Backup TCO: Shipping and Handling is Extra, As Always

Courtenay is a Channel Marketing Strategist at Zetta.net.

Tape Backup Needs Trucks
Another instance of data loss due to tape backup was in the media this weekend, with the State of California as the victim. That makes today a good time to rethink the risks and costs associated with using tape for offsite backup. To really understand the total cost of ownership for tape-based offsite backup it”™s important to consider the offsiting costs themselves, in addition to the hardware, software, and support costs.
 
In this Offsite Backup TCO calculator, where the default example datacenter environment has 20 servers and 2TB of data, using either a D2T or D2D2T architecture will run about $5,200 in offsiting costs for 1 year, growing to $20,000 over 3 years as more tapes are added to the storage vault.
 
Of course, your environment will be different, so please go ahead and use the TCO calculator to generate a custom report and let us know how accurate it is in the comments.
 
The breakdown of these costs for the example environment includes:
 
– $60.00 a week for tape pickup x 52 weeks
– $250.00 charge per emergency tape drop off, needed once a quarter
– $0.87 in monthly tape slot fees, which grow from $13 a month to $184 the first year, depending on the tape retention policy of your organization.
 
Continuing with the 20 server and 2TB example environment case, the TCO of 1 year of D2T runs about $56,000, so the offsiting costs represent less than 10% of the total cost. This sounds like a reasonable proportion, until you consider that the overall costs are 430% higher than an enterprise-grade online backup and recovery service like Zetta.
 
The cost delta is worrying, as is the tendency of tapes to go missing, break, or become corrupted. Having your data “fall off the back of a truck“ is a gut wrenching experience. The key takeaway from this data loss incident and cost analysis is that paying a premium to send critical data out to a hectic world doesn”™t make as much sense as saving money while backing it up peacefully in the cloud.

Courtenay Troxel

March 27, 2012

7 Human Errors That Cause Data Loss – And How to Avoid Them

Courtenay is a Channel Marketing Strategist at Zetta.net.

Data Loss Emergency
A recent data loss incident got us thinking about what causes organizations to lose data. It turns out that human error is the cause in 26% of data loss incidents, according to data recovery firm Kroll Ontrack. In virtualized environments, the human error rate rises to an alarming 65%.
 
The balance of data loss incidents are attributed to problems such as hardware failures. Some data loss incidents may be minor, but any loss of important business data costs the company time, customers, reputation, and revenue.
 
Here are some common backup and recovery errors and how to help prevent them:
 

1. Optimizing Only for Backup, Not Restore

Daily incremental backups are easy. They’re much smaller and faster than a full backup. However, if you’re using tapes it also means that restoring requires loading the last full backup and each of the incremental backups since then.
 
The danger comes when you need multiple backup sets to perform a full restore – the restore time increases and the number of tapes increases. Of course, the more tapes required the higher the probability that one is missing or corrupt.
 
The key here is to balance your needs for quick backups with your organization’s recovery time objective.
 

2. Only Keeping the Current Backup

When data gets corrupted, you need to be able to restore from an earlier version, before the corruption occurred. By keeping the current backup and earlier versions, you can roll back to a clean one. Some compliance requirements may also drive a longer retention history.
 

3. Not Backing Up the Right Data

Be sure to save all the essential elements of key databases, applications, or services. For example, you may be backing up your documents, but if you don’t also have the ability to restore or reinstall the application that runs them, your recovery won’t be complete.
 

4. Inadequate Storing and Transporting of Tapes

Unless tapes are sent to a reliable, experienced tape archival vendor, they can wind up getting lost or demagnetized. For example, typical fire safes are designed to prevent the contents from heating past 350 degrees, whereas safes for magnetic media like tapes must hold the temperature under 125 degrees.
 

5. Not Backing Up the Backup catalog

Make sure you backup up the catalog. If the backup catalog is inadequately protected or not backed up, a restore may require running all the tapes through the system. Or, more commonly, throwing away old backups and starting from scratch since it is impossible to accurately determine the content of the backup tapes.
 

6. Ignoring Persistent Backup Errors

Backup administrators may not notice that a particular laptop isn’t being backed up during the scheduled backup window, perhaps because it was taken home or not turned on. A comprehensive view of all systems and their backup status is needed so critical data isn’t missed.
 

7. Not Testing Restores

The biggest backup problem is not testing restores. It isn’t a valid, provable backup, until it’s been restored and verified. This applies whether you are backing up to tape, using a virtual tape library or using a colocation or disaster recovery service. To make sure the data is good before being backed up and to ensure you are backing up all the needed data, you need to test to see that it can be restored to its original condition.
 
Most of these errors can be eliminated by using a fully automated online data backup and recovery service.

December 09, 2011

7 Human Errors That Can Lose Your Backup Data “” and What to Do About Them

Jeff Whitehead is co-founder and CTO of Zetta. Previously, as CIO of Shutterfly, he successfully managed over six petabytes of storage infrastructure.Twitter: @jwhitehead

Human ErrorsAccording to data recovery firm Kroll Ontrack, human error is acknowledged as the cause for 26% of data loss incidents experienced in typical IT systems. The balance of data loss incidents are attributed to problems such as hardware failures. In virtualized environments, the human error rate rises to an alarming 65%. Some data loss incidents may be minor, but any loss of important business data costs the company time, customers, reputation, and revenue.
 
Here are some of the more common errors companies make in their backup and recovery process and steps to take to prevent such losses:
 

Doing Incremental Backups, or Optimizing Only for Backup and Not Restore

It is easy to perform daily incremental backups. They are much smaller and faster than doing a full backup. However, it also means that to do a restore requires loading the last full backup and each one of the incremental backups that have been done since that time. The danger comes when you need muliple backup sets to perform a full restore—the restore time increases and the number of tapes increase. Also, the more tapes required the higher the probability that one is missing or corrupt. You must balance your needs for quick backups with your recovery time objective.
 

Only Keeping the Current Backup, or Insufficient Version History

When data becomes corrupted, you need to be able to restore from an earlier version of the data before the corruption occurred. Don’t just keep the current backup version, but earlier ones as well, so you can roll back to a version that is clean of corruption. Compliance requirements may also drive a longer retention history.
 

Not Backing Up the Right Data, or All of the Data

If you are not backing up all your data, at least be sure to save all the essential elements of the database, application or service. For example, you may be backing up your documents, but if you also don’t have the ability to restore or reinstall the application, your recovery won’t be complete.
 

Failing to Adequately Store and Transport Tapes

Unless the tapes are sent to a reliable, experienced tape archival vendor, tapes can wind up getting lost or demagnetized. Many store tapes and other magnetic media in fire safes, thinking that protects them in the event of fire. Fire safes are designed to prevent from heating past 350 degrees, whereas safes for magnetic media must hold the temperature under 125 degrees.
 

Not Backing Up the Backup Catalog

If the backup catalog is inadequately protected or is not backed up, a restore may require running all the tapes through the system to recover the state, or more commonly throwing away old backups and starting from scratch since it is impossible to accurately determine the content of the backup tapes. Make sure you backup up the catalog.
 

Ignoring Persistent Backup Errors

Backup administrators can be overwhelmed with the volume of backups, and may not notice that a particular laptop isn’t being backed up during the scheduled backup window, perhaps because it was taken home or not turned on. A comprehensive view of all systems and their backup status is needed so critical data isn”™t missed.
 

Not Testing Restores

The biggest backup problem is not testing restores. It isn’t a valid, provable backup, until it’s been restored and verified. This applies whether you are backing up to tape, using a virtual tape library or using a colocation or disaster recovery service. To make sure the data is good before being backed up and to ensure you are backing up all the needed data, you need to test to see that it can be restored to its original condition.
 
Most of these errors can be eliminated by using a fully automated online data backup and recovery service. Such a service puts in place the necessary policies and procedures to ensure that the data is available for restoration when needed.
 
By using a service such as Zetta Data Protect, backup administrators can focus on the critical, higher value tasks of data management, instead of worrying about whether the tapes are being rotated properly. For more information, see the Zetta Online Backup and Recovery Center.

Follow Jeff Whitehead on Twitter.


Privacy & Terms | Site Map
© 2014 Zetta, Inc. All rights reserved.