Get table properties in Azure SQL Database

Even in Azure SQL Database, you need to know certain things about your tables that can be hard to find. We’ll show you how to get that data.

Even in Azure SQL Database, you need to know certain things about your tables.  For instance, you need to audit table size and row count information, as well as schemas and more.

Data not available

Recently, I found that I could only get some of the properties I wanted when I tried to get table properties from Azure SQL Database using SMO.  I could get table name, schema, and many more, but some random properties – like RowCount and DataSpaceUsed – returned NULL.

After racking my brain for about half a day, I found the problem.  Before getting into that however, let’s get into some sample code that’ll demonstrate the problem:

Simple solutions

The fix turns out to be simple: move to SQL Server 2016.  Something in previous versions of SMO prevents them from reading these properties from Azure databases.  (I’m always able to pull that data just fine from on-premises servers, though.)

But if you’re on SQL 2014, how do you get the 2016 SMO objects?  The easiest way is to download the 2016 Feature Pack.  It won’t say “SMO” anywhere, and there are a lot of packages to choose from.  What you’re looking for is SharedManagementObjects.msi.

Also, starting in SQL 2017, SMO is available as a separate NuGet package.

So even though you’ve got your data in Azure SQL Database, you don’t have to throw your hands up to all admin activity.  Many companies say they don’t need DBAs any more, because their data is in the cloud.  Nothing could be further from the truth.  And hopefully this will get you one step closer to where you need to be with monitoring those environments.

Here’s another article on the need for monitoring your environment, and not just performance.

Pay Yourself First

The biggest problem I see is the huge number of DBAs who have let themselves fall behind. Live in interview mode, and pay yourself first, and you won’t fall behind!

I’ve done a ton of interviewing in my career, both as the interviewee, and as the interviewer.  The biggest problem I see is the huge number of DBAs who have let themselves fall behind. How does this happen?

Middle of the pack

Once you’re a more or less established DBA, you have yourself a job, and you go in every day and you do your duty.  You back up what you need, do the restores you need, fix replication, look at a couple bad queries, and set up a new server for the latest project.  You’re busy as hell at work, and there’s no end in sight to all the tasks and projects and problems.

That’s the problem with being a DBA these days: you’re always swamped with problems. (The real reason is that companies are absolutely dedicated to not taking data seriously, but that’s another article entirely.)

Losing ground

So you work, and you get further and further behind the learning curve because there’s no time to do anything for yourself: no time to learn anything new, pick up a book, watch a tutorial, or even practice what you already know.  You’re always putting out fires!

Then, when it’s time to interview for a new job, you find yourself cramming in the last couple days to try to bone up on your knowledge.  Speaking as someone who’s interviewed a lot of DBAs: this definitely shows! Anyone who conducts any amount of interviews at all can tell when you’re just barely recounting something and when you know the topic cold.

Live in interview mode

Okay, I have a radical, two-step plan for your professional development. Here we go:

  1. Stop cramming for interviews like you’re trying to pass a test.
  2. Live in interview mode.

Interview mode (n.) – The practice of conducting your daily work life as if an interview could happen unexpectedly, at any time.

Take time to study every day.  It doesn’t matter how much, but I think 30 minutes isn’t too much to ask.  Even if you’re not studying to interview, your skills will get rusty when you don’t put them into practice for a while. Chances are, your company won’t ever give you the time to do it, so you have to take that time yourself.

Pay yourself first

Every day when you come in to work, take 30 minutes to work on something for you.  Learn how to do partial restores.  Learn how to set up an Availability Group.  Learn how to add a primary key to a table using T-SQL.  Learn XML, or JSON, or HTML.  It doesn’t matter.  Pick something up that you want, or something that you know you lack.

I call this paying yourself first, which is actually a financial term:

Pay yourself first is a phrase referring to the idea that investors should routinely and automatically put money into savings before spending on anything else. – InvestingAnswers.com

When it comes to your career, make sure you routinely and automatically put time into your development before spending it on anything else.

When someone comes to your desk and asks you to do something, tell them that you’re doing your daily checklist, and you’ll be with them in a few minutes.  (People at work don’t understand that daily checklists are dumb out of style, so they’ll leave you alone.)  Your company won’t give you the time to do this, so you have to take the time.

Study first thing in the morning, before things get started. Once the day really gets going, it’s hard to even remember studying, much less to find the time.

You may not be able to do it every day.  There may be some days when you walk in and there’s some emergency that honestly takes priority.  That’s okay, take care of your emergency.  But outside of that, there’s very little that can’t wait 30 minutes, especially when you’re “doing you’re checklist to make sure things are okay”.

So take some time to be good to yourself before things get crazy every day.  Improve yourself, live in interview mode, and pay yourself first.

 

 

Photo by Tim Gouw on Unsplash

Coding Naked

Let’s talk about CASE tools, and why coding for databases without one feels so….exposed.

I know the title sounds like clickbait, but it’s how I’ve felt for a long time.  Why? Because of CASE tools.

CASE stands for Computer-Aided Software Engineering. Here, I’m referring to data modeling tools like ErWin and ER/Studio.  It disheartens me that the industry as a whole has gotten away from using these tools. In this world of rushing software to market, we’ve forgotten that the data-driven marketplace starts with the word ‘data‘.

A brief history of CASE

I started in IT in the 90s when things moved a little slower and we actually took time to model databases.  [Editor: I somehow managed to resist inserting a gif of an old man and a cane.]  So, I got used to using CASE tools, and by the year 2000 I was quite adept at using the different flavors of these tools.  And I’m going to tell you, I miss them greatly.

The word domain has had a lot of play in IT, because it’s a nice generic container for many things. In the 1990s, a domain was a universal concept when working with a CASE tool. In SQL, we have another word for domain: column.  Old timers like Joe Celko get really irate when you call a domain a column, because they know what it’s for.  But we’ll get to that in a minute. First, let’s talk about the difference between a domain and a column.

Column vs Domain

A column is a single property in a table, which holds the same type of data throughout all rows.  Think of an Excel spreadsheet that stores a list of names.  That’s how data is represented to us, even in SQL Server Management Studio (and practically every other query tool out there).  We see the data as these little on-the-fly spreadsheets.  But, that would be a very myopic view of what a domain is.

In data modeling, a domain is implemented as a column. But, the domain itself is an organizational structure for column definitions.  A single domain can apply to individual columns in dozens of tables, if you like. Let’s take an example.

Let’s say you’re writing a SQL application, and you sit down to create a table for it. In that table, you need a FirstName column.  After some thought, you give that column a datatype of VARCHAR(50).  You then get pulled away, and don’t get back to the code for another week or so. When you return to the code, you pick up where you left off and start creating another table.  The second table also needs a FirstName column. As it’s been weeks, you forget that the first table has a FirstName column with VARCHAR (50), and you’re in a different mindset today, so you give this new FirstName column VARCHAR(75).

This sort of thing happens all the time, especially in projects with multiple developers. See how this app is just begging for bugs?  These bugs can be very difficult to track down.

Developing with CASE and domains

This is where domains and CASE tools come in.  In a CASE tool, you can define FirstName as a domain.  That domain a single entity where you define the datatype, length, nullability, and often even more properties.  Then, when you’re creating tables in your CASE tool, you associate each column with a domain, instead of assigning datatypes.

With a defined domain, every column you associate with that domain automatically gets consistent properties.  Furthermore, if you have a change of heart and decide to change FirstName from varchar(50) to nvarchar(100), you change it at the domain level. Then all the columns that use it pick up the changes.  (I’ve always liked how meta it is that you’re normalizing the data model process itself.)

You can see why the old-timers insist on calling it a domain: because it is.  It’s not a single instance of a FirstName…it’s the domain of FirstName, which every FirstName column will take its properties from.

Fill your data applications with domains, and not columns.  Now that you’ve got a domain for FirstName, you can create domains for every other ‘column’ in your different data projects.  Even better, you can reuse these domains in your different projects, so FirstName is the same in every application you write!  This is what we call a data dictionary.  It’s the right way to do things because it forces you to think about the right data type for the job, and then it enforces it throughout your entire data enterprise.

What’s more, publish your data dictionary! Anyone writing writing stored procedures and other code against your databases should use that data dictionary to determine what datatypes they need for parameters, variables, temporary tables, and the like.  With any luck, they would look up those values in the database anyway, so you may as well expose them all in a single, searchable place.

Coding naked, without CASE

All of this is to show now naked I feel when I code these days.  You see, I don’t have access to a CASE tool any more, because they’re so expensive.  The companies that sell them think an awful lot of the software, while the companies I’ve worked for don’t. So, I haven’t been able to talk anyone into footing the bill in a long time.

The industry as a whole doesn’t see the value in not rushing software out the door, so nobody takes time to actually model data anymore.  [Old mane cane gif, resisted again. -Editor] So, the tools should be cheaper because demand isn’t very high.

All I know is that my projects have suffered as a result.  I’ve had silly bugs and inconsistencies that are pointless, because I was in a different frame of mind when designing different tables, and didn’t realize I was creating my columns with different properties.  Until I can solve my CASE tool problem, it’s like coding naked because I’m exposed to bugs, data anomalies, performance issues due to implicit conversion issues, and probably more.

Let’s get back to a time where we can slow things down and start doing things right.  And vendors, stop treating CASE tools like a precious commodity.  Price them to own and you may see an increase in sales.

"Old man yells at cloud."
Close enough. -Ed.

Have you met Hiro?

Six things to know about Minion Enterprise: 1. You love T-SQL. 2. ME is an enterprise product. 3. No dropping objects on servers. 4. No event storming. 5. Log everything. 6. And it takes just five minutes.

Hi, I’m Hiro! You’ve probably seen me around the Internet now and again. (I do like that picture; it makes me look extra-smart.)

Okay, since we’re really, actually talking for the first time here, let me tell you a few things about me.

1. I’m a robot.

And yes, I can fly. What’s the point of being a robot-spaceship hybrid if I can’t fly?

2. I love T-SQL

I love data in general, really. But I like T-SQL specifically to get at the data. I know, the picture above shows me looking at graphics, but they had me pose with those – they’re actually these removable stickers – just for the photo shoot. You’ll almost never find me using a GUI, because T-SQL is so much more efficient. I can type, like, 1,000 wpm, but I can’t click on a GUI much faster than anyone else.

You’re a DBA, right? You know what I’m talking about.  A GUI limits what can be done with the data, anyway. T-SQL lets me, and you, and everybody, query anything at all with the data available. And let me tell you: I collect a LOT of data, and I keep it in tables. (No XML, no flat files, no proprietary formats. You’re totally welcome.)

I also put together some great stuff, like stored procedures that show you Active Directory permission chains, and alerts for low disk space, things like that.  You should hang out sometime and see all the stuff I made.

3. I like a high-level view

It probably comes from my robot superpower, which is flying.

I’ve seen that a lot of DBAs use tools that make you do a task one-at-a-time, server by server. That takes forever.  Me? I like to report and alert and manage a bunch of SQL Server instances at once. I’ve queried 10 servers at a time, and I’ve changed sp_configure options for 200 servers at once.  It’s what we at MinionWare call the set-based enterprise.

Okay, like with security. I don’t really know why you have to spend so much time on researching or scripting or cloning permissions. For me, it’s effortless. I can make your new junior DBA look exactly like the previous one, down to object level permissions, for all your servers. It’s just one T-SQL command!

4. I don’t create objects on managed servers

That’s a big pet peeve of mine. Look, if I dropped jobs or triggers or something out on managed servers, and then I needed to upgrade myself? That means the team would have to go through this big process and make sure there were plans and rollback plans and on and on. No good. This way, though, I sit on a single server, and any upgrade or change is just effortless.

5. I DO NOT LIKE event storming

One of my main jobs is to alert the team when something passes a threshold. Like, if the network goes down overnight and it messes up backups, you really need to know about that! But I think it’s frankly spammy to send dozens or hundreds of emails about it. Instead, I like smart, condensed alerts. I like to provide exceptions, adjustments, and deferments. Smart alerts. No event storms.

6. I log everything, so you can report on anything

I told you I love data, right? I follow the maxim, “log everything”. I’m really, really clever.  (I’m not egocentric; I’m a robot!) I’m clever enough to give you good reports and views, things you need. But I’m not you, so I won’t be able to come up with every clever thing you might think of.

To make up for that, I log everything. Everything I can think of that might possibly be useful, I collect and store. And I’ve spent lots of time thinking up understandable table names, so you can find the data easily. Again, you’re totally welcome.

6. I’m easygoing

It takes about five minutes to get me settled in and configured.  I don’t like to be a bother.  And once I’m in, I’ll just do my job! I mean, if you have time we can totally hang out, and I can tell you all about what I do. But if you’re busy, I’ll just watch over your systems for you, and send you alerts, and collect data you might want later for audits or disk space projection.

Seriously, I’m the most chill, hardworking coworker you’ll ever have. Download me today, and I can show you what I mean.

Reader, this is Hiro. Hiro, Reader.

Your Service Level Agreement is a Disaster

As a DBA, you’re in charge of getting systems up and running quickly in the event of an emergency. This is all right and proper, right up until you start defining SLAs. Let’s see what went wrong.

As a DBA, you’re in charge of keeping the systems healthy, and getting them back up and running quickly in the event of an emergency. This is perfectly right and proper, right up until you start defining a service level agreement.

A Service Level Agreement (SLA) defines the level of service you agree to provide, to get the system back up after a downtime. An SLA is usually expressed in terms of time. So, if you have a two-hour SLA, that means you agree that when there’s a grave issue, you’ll have the system back up within two hours.

But how did you get that two-hour SLA in the first place? Usually, it goes like this:

  1. The customer explains that they must have the database up within two hours of a problem.
  2. You don’t see any problems with that.
  3. Maybe there’s even a written agreement that you sign, knowing full well your database experience can easily handle a downed database in that time.

Like so many things, this sounds perfectly reasonable on the surface, but doesn’t hold up once it comes in contact with reality.

SLAs in the Real World

SLAs are often poorly thought-out, and very rarely tested. As a matter of fact, most companies don’t even have SLAs; they have SLOs. An SLO is a Service Level Objective – something you’re striving for, but don’t know for sure whether you can achieve. SLOs allow you to have some kind of metric, without bothering to test in advance whether the objective is even possible.

This lack of testing is the primary barrier to achievable SLAs. Lots of factors can impact your ability to get a system up and available after an issue. Let’s take a close look at just two of those factors: hardware failures, and data corruption.

Hardware Failures

When a hardware failure causes an outage, it can take a long time to get replacement parts. Companies work around this – sometimes – by keeping replacement components on hand for important systems, but that’s only sometimes. If your company has no replacements lying around, you are completely at the mercy of the hardware vendor. This can – no, will – demolish your SLA.

Now, maybe your company has an SLA with the parts vendor, and maybe they don’t. Typically, that vendor SLA will be something like a four-hour replacement window…but that’s just when they agree to show up! The vendor can’t promise that the parts will be installed, configured, and running in that time.

So, your two hour SLA won’t work with the four hours it takes just to get the hardware, and the two (or more) hours getting everything set up. Oh yes, plus the time to diagnose the issue in the first place, possibly reinstall SQL, restore the database, troubleshoot additional issues, or all the above. All of this puts you at least three times the SLA you agreed to support.

Conclusions:

  • Consider how long it takes to get replacement parts.
  • If your customer is important enough, keep replacement parts in-house.
  • Account for extra time, for troubleshooting, installation, configuration, restoring, more troubleshooting, and anything else that can come up.

Data Corruption

Database corruption is another outage scenario. Depending on the level of corruption, it may take anywhere from a few minutes to a few hours to diagnose and fix it.

For now, we’ll assume that it’s just a table that’s gotten corrupted. Now, is it the data, or just an index? If it’s an index, depending on the size of the table, it could be a quick fix, or it could be a couple hours. However, if it’s a table, then you may have to restore all or part of a database to bring it back. That brings another host of issues into the fray, like:

  • Do you have enough space to restore the database somewhere?
  • How long will it take?
  • Is the data even onsite?
  • How will you get the data back into the production table?
  • Do you bring the system down while you’re fixing it?

Of course, it could also be the entire database that’s down, in which case you will need a restore (assuming the corruption wasn’t present in the last backup). A few of the things you must consider:

  • Do you know how long that restore will take?
  • Have you done what is necessary to make sure you can restore quickly by tuning your backups, making sure the log is small, turning on IFI (Instant File Initialization), etc.?

Without some foresight, you could easily spend that two hour SLA window zeroing out 90GB of log file. Your 1.5 hours of data restore will put you quite a bit outside of your agreement.

Conclusions:

  • Make sure you have space to restore your largest database, somewhere off the production server.
  • Implement a good CheckDB solution, plus alerting.
  • Practice various recovery scenarios to see how long they take.
  • Make sure your backups are in order.
  • Practice database restores, and get your backups tuned. (Tuning your backups means you can tune your restores, too!)
  • Take these practice sessions into account when you make your SLAs.

The Solution

The conclusions above are a good start, but not at all the complete picture. If you’re keeping a single SLA for any given server, you’re doing yourself and your customers a huge disservice.

First, define separate SLAs for the different types of failure, and define what each specific failure looks like. For instance, if you define an SLA for database availability, define what ‘available’ means. Does it mean people can connect?  Does it mean that major functions are online?  Does it mean absolutely everything is online?  Does it include a performance SLA?  I’ve seen performance SLAs be included in downtime procedures because sometimes a database is so important, if the performance isn’t there, then it might as well be offline.

Next, review SLAs regularly. So, you’ve reasonably determined that you can accommodate a four-hour SLA for the DB1 database. What about, as the database grows?  Are you going to put in an allowance for the database tripling in size?  Surely you can’t be expected to hold the same SLA two years later that you did when the database was new.

Finally, test, test again, and then test one more time just to be sure. In fact, you should be testing your recovery procedures periodically so you can discover things that may go wrong, or lengthen the process. if you promised two-hour downtime and you can’t get your recovery procedures under that time, then you’ve got some re-working to do, don’t you?  Don’t just throw in the towel and say you can’t do it, because contracts may already be signed and you may have no choice but to see that it works. Maybe you’re really close to being able to hit the SLA, and you just have to be creative (and maybe, to work for a company that’s willing to spend the money).

Distinguish backup file names or pay the price!

So far, no one has found exercise to be beneficial to servers. Purposeless repetitive motion may be good for human muscles, but your SQL Server instance experiences no gain for the pain. Here’s a good example: taking useless backups. Let me explain…

Running around a track
SQL instances shouldn’t do this

So far, no one has found exercise to be beneficial to servers. Purposeless repetitive motion may be good for human muscles, but your SQL Server instance experiences no gain for the pain.

Here’s a good example: taking useless backups.

(“Did she say useless backups? I’ve never heard of such a thing!” Yeah, just wait.)

Backup file names are critical

Traditionally, backup files are named after the database and the backup type, and given a timestamp. So you’ll see something like master_FULL_20170101.bak. If you like to stripe your backups, you might name the files something like 1of5MasterFull20170101.bak, and so on.

But I have run across shops that takes backups without bothering to time stamp the file name: master_FULL.bak.  These shops either overwrite each backup file, with each successive backup (using INIT and FORMAT), or add to the backup set (which I find mildly annoying, but to each their own).

The problem with using the same backup file name over and over is if you have a cleanup mechanism that deletes old backup files!

The same-name cleanup issue

Let’s say that in your shop, you have Minion Backup (MB) installed and running with the following options:

  • INIT is enabled
  • FORMAT is enabled
  • Backup file retention (in hours) is 48, so we’ll keep 2 days’ worth of backups
  • Backup name is set to %DBName%%BackupType%.bak, which works out to DB1Full.bak for a full backup of DB1.

Note: For more information on the %DBName% and %BackupType% inline tokens, see “About: Inline Tokens“ in our help center.

Here is what happens:

  1. On day 1, MB takes a backup of DB1, to \\NAS1\SQLBackups\DB1Full.BAK.
  2. On day 2, MB takes a backup of DB1, which overwrites the file \\NAS1\SQLBackups\DB1Full.BAK.
  3. On day 3, MB takes a full backup of DB1 (which overwrites the same file). And then the delete procedure sees that it has a file from day 1 (>48 hours ago) that needs deleting. And so it deletes \\NAS1\SQLBackups\DB1Full.BAK. Remember, this is the file that MB has been overwriting again and again.
  4. On day 4, MB takes a backup of DB1, to \\NAS1\SQLBackups\DB1Full.BAK.. Then, it sees that it has a file from day 2 that needs deleting, and so deletes \\NAS1\SQLBackups\DB1Full.BAK.

See? From Day 4 on, we’re creating a backup file just to delete it again!

Fixing the issue if you have MB

One excellent way to figure out if you have this problem is to notice that, hey, you don’t have any backup files. Another indicator in Minion Backup is if you see “Complete: Deleted out-of-process. SPECIFIC ERROR:Cannot find path ‘…’ because it does not exist.” in the Minion.BackupFiles Status column.

But the real smoking gun is if you haven’t time stamped your backup files in Minion.BackupSettingsPath. Here’s how to fix that:

That will give each file its own name, and eliminate the chance of this ever happening.

And it will prevent your server from performing useless exercise. Me? I’m going to hit the gym.

SQL maintenance is a lifecycle, not an event!

You’ve heard me talk about this many times, in so many different ways, but it’s worth repeating: SQL maintenance lifecycles are important.

People who disagree, disagree because they spend too much time firefighting and don’t have time to really think about it. Don’t know anything about SQL maintenance lifecycles? Let’s touch on a couple of points…today, we’ll cover SQL backup.

You’ve heard me talk about this many times, in so many different ways, but it’s worth repeating: SQL maintenance lifecycles are important.

People who disagree, disagree because they spend too much time firefighting and don’t have time to really think about it.  Don’t know anything about SQL maintenance lifecycles? Let’s touch on a couple of points…today, we’ll cover SQL backup.

SQL maintenance lifecycles: Backups

Backups have the biggest lifecycles of all, because the uses for backup files are so varied.

They’re used for restoring databases to the original server, sure.  But here’s a short list of the other kinds of things you need to do with backup files:

  • Restore to development, QA, etc.
  • Send to vendor
  • Initialize replication
  • Initialize Availability Groups
  • Restore to a secondary server to run CheckDB

A backup is not a single event

A database backup isn’t a single event.  But unfortunately, nearly every vendor out there – and most DBAs – treat backups as a single event that has a start and a finish.  “Look, I took a SQL backup and produced a file, and now I’m finished!”  But that’s not remotely the case.  We use and shuffle SQL backup files around all day and all week long on our various servers.

This is why I’m so against those tools that I call “network backup tools”.  You know the ones: Backup Exec, AppAssure, etc.  Your network admin loves those tools, because he doesn’t understand SQL and he wants the same backup utility for the whole shop.

But network backup tools simply aren’t going to cut it for SQL maintenance lifecycles, because they take snapshots of the data and log files.  And quite often those two files aren’t fully synchronized when the snapshot takes place, so you wind up creating a new log file and losing whatever data was in there.  Not only that, but you can only attach those files somewhere, usually on the same server they came from.

Without the understanding that your maintenance has a lifecycle, you wind up with a lot of piecemeal solutions.  You write something to restore a database on a development box, then another DBA does the same thing, then another, then another.  Soon you have a huge mix of solutions that need support, all created by a mix of DBAs over the years…it’s an unmanageable web of solutions, each one with its own quirks and limitations.  That’s no way to run a shop, and it’s no way to support your data.

SQL Maintenance Lifecycles: SQL backup done right

Minion BackupThis is why we’ve made Minion Backup the way we have.  Minion Backup is the only backup tool on the market that treats your environment holistically.  Minion Backup understands your SQL maintenance lifecycle and it works for you to provide you built-in solutions that you just deploy.  There’s no need to write your own processes, or manage something separate because it’s built into Minion Backup by default.

Let’s take for instance that you need to back up your SQL certificates, because you know, who doesn’t?  With Minion Backup you can easily backup every certificate, even to multiple locations.  In fact, you can back up your certificates to as many locations as you need.  This functionality is built into Minion Backup, and setting it up is a simple table setting.

Not only that, but since we store the private key password encrypted, and give it to you in the log, you can even rotate the private key passwords on any schedule you like to increase your security.  This way if you change the private key password every month, then you’ve always got access to the encrypted value in the log so you can restore that certificate with ease.  This is what we mean by SQL maintenance lifecycle management.  All of our SQL maintenance products have this level of lifecycle management.  And we’re improving them all the time.  We work closely with our users to give them built-in solutions that matter.

All of our SQL maintenance products are completely FREE, so download them and try them now.  And if you like what you see, then try our enterprise products and see the power of bringing your SQL maintenance lifecycles into the enterprise architecture.

Download our FREE maintenance modules below:

  • Minion Backup: Free SQL Server Backup utility with world-class enterprise features and full lifecycle management.
  • Minion Reindex: Free SQL Server Reindex utility with unmatched configurability.
  • Minion CheckDB: Free SQL Server CHECKDB that solves all of your biggest CHECKDB problems.

 

Integrity Checks are Useless*

Integrity checks have become more and more important over the years, as companies store more and more data. This data is absolutely critical, so when corruption happens, it can break a company. But in all but the smallest shops, running any kind of integrity check is all but completely useless. Unless you have Minion CheckDB.

Integrity checks have become more and more important over the years, as companies store more and more data. This data is absolutely critical, so when corruption happens, it can break a company.

But in all but the smallest shops, running any kind of integrity check is all but completely useless.

Integrity checks without management

DBCC CHECKDB was originally written without consideration of management. For a very long time, you could only output CHECKDB results to the client – meaning you had to run it by hand – or to the error log. Seriously, what good is that? Later, Microsoft added the ability to output CHECKDB results to a table, but that remained undocumented for years so very few people knew it even existed.

That’s not the only problem. A lot of DBAs don’t run CHECKDB at all, because large databases take a long time to process. CHECKDB takes an internal snapshot for its operations. For large databases, that snapshot can fill up the drive, killing the integrity check process entirely. So, you have nothing to show for your efforts. Some DBAs tried to get around this issue by moving to CHECKTABLE instead, but that process comes with the same issues, and is even harder to manage because there are so many more tables than databases.

Integrity checks without reporting

Another reason that running DBCC CHECKDB is useless is this: reporting is still very difficult. Let’s say you have 50 servers – yes, I know a lot of you may have many more than that – and CHECKDB is set to run through a maintenance plan, outputting results to the SQL error log. In this scenario, you have to mine the log for CHECKDB results. That means you not only have to know when the CHECKDB completes on each and every server, but you also have to know what the log entry will look like, and how to parse out the results so they’re readable…again, on every single one of 50 servers. There is simply no easy way to consume CHECKDB results on any kind of scale.

Useful integrity checks with Minion CheckDBMinion makes integrity checks useful again

The whole goal of Minion CHECKDB was to transform these issues surrounding integrity check processing into simple settings. Now DBAs of every skill level can have a stable, standardized integrity check process for even their largest databases.

Minion CHECKDB has a huge number of groundbreaking features. MC solves the reporting problem by capturing integrity check results to our own log table, along with a huge amount of additional information. Now it just takes a simple query to get the results. Instead of the native raw CHECKDB results, the MC log table tells you how many corruption errors were encountered. Then you can drill deeper from there if you like.

We’ve also solved the problem of the large database that can’t be processed because it runs out of snapshot space. In fact, we’ve solved that a few ways, and I’ll discuss a couple of them here.

  • You can create a custom snapshot on a different drive that has more space. MC manages those snapshots for you, so it’s literally a set and forget operation.
  • Or, you can restore a recent backup of your database to a remote system and run CHECKDB against that instead. Remote CHECKDB even brings the results back to the primary server, so you don’t have to go hunting for them. Again, this is a setting; you configure it, and then MC manages the whole process for you.

Of course, there’s more

There are many more features in Minion CHECKDB that I didn’t mention. But since MC is completely free, you can download it and install it on as many instances as you want. Go explore!

Put MC on all your systems and start doing integrity checks again. We’ve taken away all the excuses. And when coupled with Minion Backup and Minion Reindex (both also free), you’ve got a maintenance suite that meets all your needs from planning to reporting, and everything in between.

 

*Without proper management and reporting!

Managing Backup Issues in Availability Groups

Let’s talk about some of the challenges that DBAs face when managing backups with Availability Groups. Then we’ll talk about solving these issues with Minion Backup.

A user asked me the other day how to go about managing backup issues in Availability Groups.  Well, Minion Backup is fully Availability Group aware, and has innovative features that makes it a complete solution.  This is one of the big reasons we coded MB: any solution for dealing with the issues listed below has to be manually coded. We wanted to stop re-inventing the wheel over and over.

Let’s talk about some of the challenges that DBAs face when managing backups with Availability Groups. Then we’ll talk about solving these issues with Minion Backup.

Managing Backup Issues in Availability Groups – The Issues

1. Back up from…wherever

Availability Groups don’t let you pick a specific server to run backups on.  You can only give weight to a specific node, but you can’t say that you want logs to run on Server1 and Fulls to run on Server2, etc.  You can code those choices in, but if you decide to make a change, you have to make it on all the nodes manually.

2. Mixed file locations

Pretty much every backup solution on the market appends the server and/or instance name to the backup path.  Usually, this is a good idea: if you have a centralized backup drive, you’ll want your backups separated by server name.  Availability Groups introduce a problem though, because you can back up on different nodes. So of course, each backup will be under a different path (based on the node name).  Full backups could be going to \\NAS1\SQLBackups\Server1, while the log backups for the same Availability Group would be going to \\NAS1\SQLBackups\Server3 and even \\NAS1\SQLBackups\Server2.  The backups can move around as nodes go up and down.  Restoring in this scenario can be difficult if all the backups aren’t in one place.

3. COPY_ONLY woes

Backups have to be run with COPY_ONLY on non-primary nodes.  This isn’t an issue on its own, but it is something that you must code into your solution.  And if you remove a database from an Availability Group, then you have to remember to manage that in your code (or preferably, to have coded it properly in the first place).  Either way, it’s just one more thing to manage.

4. Work multiplier

Any change to a schedule, or to a backup setting must be made on all nodes.  Nothing in the Availability Group setup lets you manage your multiple backup locations, and remembering to do this on all the nodes is doomed to fail.

5. Backup file backlog

If your backup node fails, you lose your record of where its backups were. Without that log, the backup files won’t delete on your defined schedule.  Sure, you could probably get that log back by restoring a copy of the management database from that server, but how much time has gone by since then?  That is, of course, if the management database was being backed up properly.  And if all that happens, it’s just one extra step you don’t need when you’ve got a node down.  You shouldn’t have to worry about those tasks that should be taken care of for you.

These are the biggest issues with managing backups in Availability Groups. Now let’s look at Minion Backup. It solves each one of them with no coding required.

Managing Backup Issues in Availability Groups – The Solutions

Availability Groups: Managing Backup Issues

1. Back up from the right server

Backing up from a specific server is easy in Minion Backup.  Minion Backup knows if your database is in an Availability Group, and you can set the “PreferredServer” column in the Minion.BackupSettingsDB table to tell it where you want each backup type taken.  You have the option to use the backup preference in the Availability Group settings by configuring PreferredServer to ‘AGPreferred’, or you can use the name of a specific server.  That means you can set a specific server for fulls, and a specific server for logs if you want.  If you want Minion Backup to determine your PreferredServer dynamically, you can code your own logic use it as batch precode.  While it’s an easy setting, you can make it as flexible as you need it to be.

2. Files in one place

We know that backups coming from different servers means that all of your different backups can be in different locations.  However, Minion Backup solves this problem easily using server labels.  The ServerLabel field (in Minion.BackupSettingsPath) overrides the ServerName in the backup path, so that all of your backups go to the same location.  You can set ServerLabel to the Availability Group or listener name, or even the application name if you want.  So, instead of sending backups to \\NAS1\SQLBackups\Server1, the backups would go to \\NAS1\SQLBackups\MyAGName.  As long as you set all the nodes to the same ServerLabel, then no matter which node they backup to, they’ll all go to the same location.  Now you don’t have to search around for where the different backups are and your restores and file deletes are easy.

3. COPY_ONLY automatically

If you want to take a full backup on a non-primary node of an Availability Group, then you have to use COPY_ONLY.  And of course, you do have to code that into your own solution.  However, Minion Backup is fully aware of the Availability Group, so it does this for you automatically.  There’s nothing you need to code, and nothing you need to manage.  It just does it.  As databases come and go out of your Availability Groups, Minion Backup makes the right choice and does what it needs to do.

4. Sync settings between nodes

Manually changing settings on each node of your Availability Group is just asking for trouble.  You could have a new node that you haven’t heard about, or you could just forget about one, or a node could be down and then you get busy and forget to push the changes when it comes back up.  So what you need is a way to automatically push all of your settings changes to all the nodes.  This is where Minion Backup comes in.  All you need to do is turn on the SyncSettings column in the BackupSettingsServer table and Minion Backup will keep all your node settings in sync.  This also means that as nodes come and go, they’ll get the new settings, and if a node is offline, Minion Backup will push them when it comes back online.

5. Backup files, handled

If you could keep your file deletion schedules even when a node goes down, or to be able to easily find the file locations, it can keep you from having a bad day.  Minion Backup spares you the worry about where your files, are or whether they’re going to still be deleted on schedule.  Again, Minion Backup makes this very easy with the SyncLogs column in the BackupSettingsServer table.  Your backup logs will automatically be synchronized to all nodes, so locations and delete times are always in place.

That’s a good roundup of the Availability Group features in Minion Backup.  We love these features, because we’ve seen firsthand how much easier it is to manage your Availability Group backups with Minion.

Go on, solve your problems: download Minion Backup!

Database snapshots for everyone!

If you’re lucky enough to have SQL Server 2016, go on and upgrade to SP1….you get extra-lucky special database snapshots in any version! Oh what’s that? You don’t use snapshots? Well let’s see a couple of good use cases for them.

Database snapshots have been available in SQL Server Enterprise edition in the last several versions. But if you’re lucky enough to have 2016 at work, go on and upgrade to SP1….you get extra-lucky special database snapshots in any version!

(Be sure to read the footnote for “Database snapshots” in that link.)

What are database snapshots for?

Oh what’s that? You don’t use snapshots? Well let’s see a couple of good use cases for them:

  • Take a database snapshot before an upgrade, data load, or major change. It’s WAY faster than a backup, and you can roll back the changes if you need to, almost instantly.
  • Customize integrity checks. When you run DBCC CheckDB or DBCC CheckTable, behind the scenes SQL Server creates a snapshot of the database to run the operation against. You can also choose to create a custom snapshot…among other reasons, so you can specify where the snapshot files are placed. This is especially useful for shops running tight on disk space.

Speaking of integrity checks and snapshots, Minion CheckDB is slated to come out on  this coming February 1.  Be on the lookout…it’s going to blow your mind.

Sign up for your trial of Minion Enterprise today, and sign up for our newsletter for updates!

productimg_checkdb