I’ve created some simple scripts for creating subscriptions and their pull jobs on single or multiple subscribers. It makes life much easier to use scripts, especially if you’re bringing in multiple servers and / or the subscribers are running SQL Server Web edition as they don’t get created automatically via the GUI; which can be confusing.
There are various tools to run scripts across multiple servers, my favourites are SQL Server Central Management Servers or if you want something more advanced and easier to use, try SQL Multi Script from Redgate.
You can also use Aireforge Studio to compare the jobs afterwards, making sure they were created properly and are running. It’s also useful for flagging up other configuration inconsistencies such as missing database objects, users or database and instance configuration settings.
Create local subscriptions and generate publisher scripts
Create the local subscriptions and generate the script for adding the subscriber to the publisher (this will need to be copied out and run against the publisher database).
Create and start the replication agent jobs
Remember to locate and set the location of DISTRIB.EXE. I’ve also included an option to set an alternative snapshot location. Very useful for remote servers, especially if you distribute the snapshot files using DFS-R (search here until I blog about it).
Now check that the agent is working properly and either waiting for the initial snapshot to be created or propagating out any active snapshots.
Please let me know if you spot any issues or would like to improve the scripts.
I’m writing an article covering common SQL Server configuration issues and as I write about how Microsoft have addressed some of these in SQL Server 2016, I started thinking about what they’ve missed or what they should include in the next release, be it SQL Server 2018 or version 14.
RCSI as the default isolation level
Read Committed Snapshot Isolation Level should really be the default for all new SQL Server databases and Microsoft should encourage this by changing the default database option for Is Read Committed Snapshot On to true. Upgraded databases would continue to have this set as false and new databases would also have the option to use the old version if required.
In the old days we had our own database maintenance scripts and now most people just use Ola Hallengren scripts (because they’re awesome) but these should have been shipping with SQL Server for years. Even if it’s a maintenance plan (rather not), SQL Server should ship with cross database jobs that handle index maintenance and integrity checks automatically. Either using a default schedule or a configured schedule during the installation.
Increase Cost Threshold for Parallelism
The cost threshold for parallelism should be updated to reflect modern hardware. I’m not sure how it came to be 5, but I understand it’s a number based on late nineties hardware. I generally set this to a minimum of 30, sometimes much higher but 30 would be a much better default than 5. A better solution would be to change this figure to represent seconds or milliseconds as it once did.
Format drives during install
Most SQL Server installations have dedicated drives for the data, logs and tempdb but they’re not always formatted correctly, with many systems still using an allocation unit size of 4KB, rather than 64KB. It would be useful if the installer formatted drives but a simple check of the allocation unit size and a warning would be helpful.
Server memory as a percentage
I work with the rule that a SQL Server instance should use 90% of the available RAM, unless it’s on a small system with not much to play with. On shared servers you may reduce this percentage further, but in most cases the figures your inputting are based on a percentage.
This will also fix the problem where the available RAM has changed but SQL Sever hasn’t been updated. This often happens after infrastructure upgrades where available and much needed RAM is going to waste or virtual machines that have been resized to 32GB but SQL Server is still configured to use 58GB.
Compress backups as default
It’s very rare that people would not want to compress their backups, therefore this option should be enabled by default. Saving space and time for those that backup over the network.
Aireforge OmniCompare 0.9.0 was the first release to ship with the CLI module. This is a separate application, allowing comparisons to be called from 3rd party applications, such as SQL Server Agent jobs, custom scripts and software deployment tools including Octopus Deploy. OmniCompare CLI will then feedback basic information using termination or error codes; much like log shipping alerts.
The information required to perform the comparison and the credentials required to access the servers are stored within the .adc export file, therefore the main OmniCompare application will be required to perform any future changes such as new servers or new / updated comparisons.
What is OmniCompare?
OmniCompare performs a growing list of comparisons, which can be restricted if certain values are known to differ, such as core count or edition (e.g. Enterprise vs Developer). All comparisons in OmniCompare are available in OmniCompare CLI.
A common use case for OmniCompare is checking production against staging and staging against test. This manual task could be automated using a SQL Server Agent job or building OmniCompare into your deployment process, checking for differences and aborting the release if any are found.
Comparing trace flags via the command-line
Creating the comparison file
Open OmniCompare and select the servers you wish to compare. Once selected, you can now choose a full comparison or restrict it to individual checks.
For the purposes of this blog, we’ll limit the comparison to trace flags.
Note: Although trace flag comparisons work for both SQL Server and AzureDB, many of them are not interchangeable as AzureDB is a very different database to SQL Server. AzureDB also has many trace flags enabled by default (17 at the time of writing).
Checking current values
If no differences are found, nothing will be displayed. Although it’s often interesting to check the current values. This can be done by changing the results filter from Differences only to All.
Showing all results has highlighted that each instance has the non-standard trace flag 1222 enabled, which enabled by Redgate SQL Monitor to improve deadlock information.
Exporting the comparison file
OmniCompare CLI is a lightweight application, therefore the comparison file must be created using the main OmniCompare application and exported. This exported file contains the servers to check, authentication credentials and the comparisons to perform.
To create the .adc file, select the command line icon at the top right of the comparisons section. You will then be prompted to enter a secure password and the location of the saved file.
The OmniCompare CLI executable resides in the OmniCompare program folder. The argument for the application will be logged out if none are passed in or the help argument (-h) is use.
Performing a CLI comparison
To perform a trace flag comparison against the servers detailed above we simply pass in the comparison file and password.
For the purposes of this blog I have enabled full logging, although this should be used for debugging only as a lot of information is logged; especially during full comparisons of large estates.
omnicomparecli.exe -i d:\tf.adc -p <password> -v full
This command could then be run from a SQL Server Agent job or from software deployment tools or from your own custom scripts.
We will now enable trace flag 1117 on the SQL2014 instance to demonstrate how the application changes. OmniCompare CLI has highlighted the change and returned an error code of 1. This error code could be caught and an alert raised or a deployment aborted.
The information detailed above could be logged to file using the -o command, although a simpler option would be to run the comparison via the OmniCompare application.
Discrepancies between configuration settings or database objects can be difficult to identify, especially with larger estates or multiple developer environments. Custom scripts or policy-based management could be used to check some of these settings but they will require updating and only work if the script or rule has been created.
OmniCompare checks a large and growing number of settings, highlighting unexpected changes and preventing configuration drift between instances that should be the same.
Other useful pre-deployment checks
Counts of database objects (tables, columns, indexes, functions, views etc.).
Operators, users or permissions have been applied,
and database settings such as delayed durability, RCSI etc.
So we talked a lot about 0.8.3 and how great it would be but then skipped it and went straight to 0.9.0. First of all, apologies for promising an upgrade and then delaying it but we had good reasons for this.
Version 0.8.3 was supposed to include a couple of new features (mainly CLI) and bug fixes but during our design discussions we decided to completely rewrite how OmniCompare works internally, along with changing most of our development processes. These changes now enable us to easily introduce some really exciting features for 1.0.0 and even 2.0.0 (big plans here). They also significantly reduce our development time so we can focus our attention on new features and even new products (OmniGuard being one).
I won’t list every new feature as these are listed in the release notes but the main ones are:
Improved UI for all screens.
They’re now simpler to use, support DPI scaling (for those lucky enough to have 4k monitors) and it just looks better.
OmniCompare CLI (separate blog post coming soon).
Bug fixes around connections and improvements to the connections dialog.
Version specific comparisons.
We now support major versions but we’ll support minor version changes soon. This is important because MSFT updated a number of DMV’s in 2008 and 2008 R2 during service pack updates and also added some useful ones too.
Tons of new comparisons including users, passwords, OS information, trace flags (really interesting) and more AzureDB comparisons.
Lots of changes in the background to help with new features and development.
Updated the website with changes to the download / update process.
Finally, we now have a Wantoo Ideas Board which enables the community to suggest or vote for new features (any feedback is welcome).
We also have a busy month ahead of us with the 0.9.1 release next week (will happen this time!) and some exciting new integrations with the makers of the worlds most trusted database tools.
If you haven’t downloaded OmniCompare yet you can do so here or update by simply opening the application (assuming you have an Internet connection). Thanks for reading and taking an interest in what we’re doing,