Did you Test Your Disaster Recovery Plan?
Tomorrow's Expo
FREE SSWUG Expo - Maximizing SQL Server Uptime - Tomorrow 9am Pacific
SQL Server uptime is a key component of managing your SQL Server(s). Knowing what's important, what options you have for building out solutions to help when things go wrong, is at the core of the responsibilities for those working with databases. This expo will feature information moving to SQL Server 2008 to get you started. You'll learn about new features and important things that you'll want to know about as you work with SQL Server 2008. We'll also have segments on Transactional Replication for High Availability, Database Mirroring and even log shipping basics. Join us to learn about the key elements to making sure your systems are able to provide excellent uptime options for your business, clients and end-users. Sign up today. <- click to register...
Did you Test Your Disaster Recovery Plan?
As I read responses from our SSWUG readers regarding their sever room washing, it occured to me that they had good disaster recovery plans.
Most of us are good at putting together a disaster recovery plan for our servers. We have backups taken offsite, replication, mirroring, log shipping, co-locaiton, even SAN replication. Those are just a few of our recovery techniques. There are even third party tools that can keep you alive.
I wonder how many of the disaster recovery plans for companies impacted by 911 actually worked? I'll bet someone out there has tried to gather some statistics. There are a lot of pieces to recovering from a disaster...and they may all be different.
What if you have to replace hardware and restore from tape? Some backup software is smart enough to ignore things that may be different from the previous machine to the one replacing it. But what about having to purchase a different hardware configuration? If you are working with servers, can you readily point your applicaitons to new servers? Did you try it? How long does it take?
Most big companies take this as a matter of course. But some of us smaller businesses forget that disasters do happen, and a trail run is the only way to make sure that our plan really works.
Got a story to share...send it to btaylor@sswug.org.
Cheers,
Ben
Featured Article(s)
Surface Area Configuration Tool (part 2 of 3)
In this part we will talk about all about full text service and how to use them wisely. Then we will discuss exactly what CLR is and how DAC can help when your server is unresponsive. We will talk about what SQL mail's replacement is and what database mail is.
Featured Script
CheckAllTables - dbcc on all tables
Runs DBCC against all tables in the database. ... (read more)
|
7/29/2010
|
Wash Your Servers...Who Knew?
SSWUG July 30th Expo
Twomore day to sign up for the FREE SSWUG Expo - Maximizing SQL Server Uptime - July 30th
There are so many methods for keeping your database online. Each has it's benefits and associated costs. What's the best solution for you? Come reveiw the options that are provided in SQL Server 2008. When you attend the SSWUG Free expo you'll learn about new features and important things that you'll want to know about as you work with SQL Server 2008. We'll also have segments on Transactional Replication for High Availability, Database Mirroring and even log shipping basics. Join us to learn about the key elements to making sure your systems are able to provide excellent uptime options for your business, clients and end-users. Sign up today.
Wash Your Servers...Who Knew?
Ok...I thought water in the server room was a bit of a unique event. Well, I guess not. Three of our SSWUG readers respond with their own personal experiences. This is a testament to good Disaster Recovery Procedures.
Mark Tells us...
While working for a global engineering corporation I received e-mails from the servers that they were all down on a Sunday night at 11pm and would not reboot. I went into the office and heard water running. As I walked down the hall I saw water seeping out from under the wall to the server room. I opened the door and saw water running down from the ceiling over the server cabinet and through it over every piece of hardware inside. It looked like a dishwasher with a glass door.
I shut off all the power and went to look for the main water shut off. I called in the team and we took the rack apart. We lost 2 switches, power supply in one switch, the power supply in the array, one server and the motherboard in another server. We took everything apart to include the array and dried everything with fans and dehumidifiers. Dell sent a tech out to replace motherboards and power supplies. We did not lose any data whatsoever. We were back up and running at 100% by 4pm on Monday afternoon.
Tom remembers...
Water in the server room? How about a flood of your city? New Orleans, 2005: Katrina hits, Army Corp of Engineers’ levee design flaws manifest themselves, flood ensues. Our company’s offices are at the top of a 13 story building and dry as a bone, but our whole city is broken. Two weeks and some new server shipments later, I’m rebuilding our databases and infrastructure from our backup tapes on new servers in our new home in Milwaukee. Turns out the flood/relocation was a blessing in disguise, business is great.
Bob writes...
Had the same problem one day. Got a call from the maintenance person one night: We got 2 inches of water in the server room. Rushed there and shut off equipment.
A water pipe in the attic over the server room froze and broke. It got two servers and a tape drive unit. All three had to be replaced.
Featured Article(s)
Tips for using SQL Server 2008 bulk copy
In this article, you can find some tips to use SQL Server 2008 bulk copy.
|
7/28/2010
|
SSD - One Last Time
Webcast: Understanding the new culture of SharePoint
You have a sponsor for SharePoint development, you have a budget, now where are all the users? Online Collaboration powered by SharePoint is not only a new technology, but a new culture of doing work. In this session we'll explore change management (the fluffy kind) strategies that can be tailored to your user base. Attract users, and keep your budget!
Presented by: Adam Levithan
> Register Now
> Live date: 7/28/2010 at 12:00 Pacific
SSWUG July 30th Expo
FREE SSWUG Expo - Maximizing SQL Server Uptime - July 30th
There are so many methods for keeping your database online. Each has it's benefits and associated costs. What's the best solution for you? Come reveiw the options that are provided in SQL Server 2008. When you attend the SSWUG Free expo you'll learn about new features and important things that you'll want to know about as you work with SQL Server 2008. We'll also have segments on Transactional Replication for High Availability, Database Mirroring and even log shipping basics. Join us to learn about the key elements to making sure your systems are able to provide excellent uptime options for your business, clients and end-users. Sign up today.
SSD - One Last Time
A Frequent SSWUG participant, Jeremy Lowell shares some real life experience with SSD...
I’ve had the opportunity to do some scaling for a client where we put four 320GB fusion-IO cards to the test. This database consisted of many files, partitions etc… The purpose of the exercise was to see how far the db design could scale and what hardware it would take to get it there.
We used many servers for this test but the Dell R910 was the primary DB server. Hexa-core chipset (E7540)*4, 128gb of ram etc… We utilized a netapp 3170 SAN.
Early on in the scalability testing, we discovered that our bottle neck was I/O (an expected result). We already had the fusion-io card’s ready to go so we began by offloading tempdb to one fusion i/o card. This eliminated a lot of contention from the SAN; but we still couldn’t hit the transactional throughput that I believed possible on that hardware. So, we offloaded the .ldf file and the improvement was significantly greater than I thought it would have been.
My unproven, untested reason for the improvement revolves around the WAFL technology that Netapp employs. I’m not certain that contiguous writes perform in a linear fashion. At any rate, we saw transactional throughput rates 40 to 50 percent greater after we had moved tempdb (data and log) and the databases log file off of the SAN and to the fusion-io cards.
The incredible thing about this is that up to 50 percent (HALF) of the overall I/O created from this db was able to be handled by TWO fusion-IO cards. 2. Incredible. If you need some serious I/O throughput and the storage size isn’t a huge consideration (tempdb, log); SSD can provide a lot of bang for the buck.
The comparison is having another SAN capable of handling the IOP requests (regardless of storage size) and while the cost of the IO cards is very high from a storage capacity perspective, they are dirt cheap when it comes to I/O throughput.
Well, unless someone who has really tried some stuff has more to offer, I think we'll move on to a new topic tomorrow. Drop me a note with your comments, experience or concerns at btaylor@sswug.org.
Cheers,
Ben
Featured Article(s)
Creating Databases (Part 3 of 4)
In this part we will talk about auto grow. We will also talk about the many many auto options available.
Featured Script
dba3_fx_BusinessDaysInRangeFromvBusinessCalendar_Article
Modeling Date Logic III: Implementing Business Calendars (by example in MS Sql Server 2000)ttp://www.sswug.org/see/17379 Demo... (read more)
|
7/27/2010
|
SSD - Reloaded
Webcast: Understanding the new culture of SharePoint
You have a sponsor for SharePoint development, you have a budget, now where are all the users? Online Collaboration powered by SharePoint is not only a new technology, but a new culture of doing work. In this session we'll explore change management (the fluffy kind) strategies that can be tailored to your user base. Attract users, and keep your budget!
Presented by: Adam Levithan
> Register Now
> Live date: 7/28/2010 at 12:00 Pacific
FREE SSWUG Expo - Maximizing SQL Server Uptime - Friday 9am Pacific
SQL Server uptime is a key component of managing your SQL Server(s). Knowing what's important, what options you have for building out solutions to help when things go wrong, is at the core of the responsibilities for those working with databases. This expo will feature information moving to SQL Server 2008 to get you started. You'll learn about new features and important things that you'll want to know about as you work with SQL Server 2008. We'll also have segments on Transactional Replication for High Availability, Database Mirroring and even log shipping basics. Join us to learn about the key elements to making sure your systems are able to provide excellent uptime options for your business, clients and end-users. Sign up today. <- click to register...
SSD - Reloaded
Before you get completely bored with SSD I received a couple responses I thought you may want to take into consideration.
Brice reminds us that like any hardware device you need to be sure to have a backup and restore strategy. He writes, "I was burned by a ### 128GB (SSD), because I didn't backup my pst's. Lesson learned. It lasted < 3 months! (of pretty heavy use)."
I appreciate the email heading from Brice, "Burned by SSD (My Own Fault)." We all know that any hardware can fail and eventually will. We have to prepare for that eventuality. SSD is a lot newer technology and is not as heavily vetted as the traditional disk drive. So, be sure to have a failover plan.
If you want to review some of your database failover options, SSWUG is hosting a FREE Expo on July 30th titled Maximizing SQL Server Uptime. This will be a great opportunity if you find yourself responsible for your database servers and don't yet have a strategy for system failures.
Ron is doing something different than most database applications. He is using SSD to help optimize his embedded applications for real time testing. He has found that the additional performance of SSD allows his testing tools to be much more robust and extensible.
These responses got me wondering. Could I get an inexpensive SSD for highly utilized database objects and increase my overall server performance? What if I could put something together with a raid controller using 2.5" SATA II SSDs. You can get 160G for $400 or less. Wouldn't that scream on a raid 10? You hardware gurus probably know...I wouldn't guess the SATA controllers are fast enough to compete with SAN speeds. But there are SCSI controllers out there that can work with SATA drives.
I could see putting TempDB or even SWAP files onto this kind of drive. TempDB is a great candidate. If the drive fails move it to another drive and you're back in business.
Drop me a note with your comments, experience or concerns at btaylor@sswug.org.
Cheers,
Ben
Featured Article(s)
DBA Learning Strategies (Part 3 of 4)
Next we need to document what you need to know... We found out what we know, we have our reds and greens. Now we need to figure out, which of these reds do I care about?
|
7/26/2010
|
Water Cool Your Server Room
Webcast: Understanding the new culture of SharePoint
You have a sponsor for SharePoint development, you have a budget, now where are all the users? Online Collaboration powered by SharePoint is not only a new technology, but a new culture of doing work. In this session we'll explore change management (the fluffy kind) strategies that can be tailored to your user base. Attract users, and keep your budget!
Presented by: Adam Levithan
> Register Now
> Live date: 7/28/2010 at 12:00 Pacific
SSWUG July 30th Expo
FREE SSWUG Expo - Maximizing SQL Server Uptime - July 30th
SQL Server uptime is a key component of managing your SQL Server(s). Knowing what's important, what options you have for building out solutions to help when things go wrong, is at the core of the responsibilities for those working with databases. This expo will feature information moving to SQL Server 2008 to get you started. You'll learn about new features and important things that you'll want to know about as you work with SQL Server 2008. We'll also have segments on Transactional Replication for High Availability, Database Mirroring and even log shipping basics. Join us to learn about the key elements to making sure your systems are able to provide excellent uptime options for your business, clients and end-users. Sign up today. <- click to register...
Water Cool Your Server Room
I was talking with a colleague last weekend. He was off work for two days because his server room was being dried out; literally.
They were having repairs done on the roof of their building. A water pipe broke and the water sprayed all over in the room below. You guessed it, the server room. Every server, switch, the SAN, you name it was soaked in water. They fixed the leak, fixed the roof and waited 2 days for everything to dry out. Then they brought everything back Online with no problems.
When he told me the story I said, "Wow, you guys didn't lose anything?" His reply took me off guard. They lost a great deal, the cost of which is still being calculated, simply because all of their systems were offline for two days. Everyone was reduced to their individual workstation. Printing was often impossible because the switches were turned off. Developers were unable to access version control. Sales couldn't get to their Email. There was no Internet access. Their web sites were completely offline.
I'm sharing this because I found it interesting. The fact that your server can take a bath and still work amazes me. At the same time, I think the scenario demonstrates that its very difficult to prepare for all catastrophes. This company was well financed and still brought to it's knees.
You don't always have problems at this scale...sometimes the issues may be at the server level. Perhaps now is a good time to schedule in your calendar to attend the SSWUG July 30th Expo on Maximizing SQL Server Uptime. Who knows when someone will pour a bucket of water on your production database server?
Got a disaster that you'd like to share with our SSWUG readers? Send it to me at btaylor@sswug.org.
Cheers,
Ben
Featured Article(s)
Next Generation Data Warehousing in SQL Server 2008 (Part 1 of 2)
In this session we will cover next generation warehousing in SQL Server 2008. We will cover what things are new. We will talk about what we did with data compressions and back up compressions. We will also talk a little about partition table parallelism, bettering star join queries, and star queries in the real world.
|
7/23/2010
|
|
