Infrastructure at your Service
I made a mistake recently when adding a new node to an existing Windows Failover Cluster supporting AlwaysOn SQL Server instances. When adding a new node using the GUI there’s this option checked by default named “Add all eligible storage to the cluster”. Before performing the operation the disks on the 2 nodes of my cluster were used by the SQL Server instances. Because I’m using the AlwaysOn features the disks are local on…
Introduction Sometimes you face interesting challenges with unusual environment. One of my customer needed a automated and flexible backup solution. Said like that nothing very complex you will say. But if I mention that some databases were 60TB big with more than 30 filegroups and around 600 database data files each and moreover synchronized in an AlwayOn availability group, it is not the same story and you can easily imagine that working with standard backup…
As you probably know Microsoft announced an official name for SQL Server vNext during the last Microsoft Data Amp event on April 19 2017. It becomes officially SQL Serve 2017. In my first blog post, I wrote about SQL Server on Linux and the introduction of availability groups features. At this moment the SQL Server release version was CTP 1.3. As a reminder, with previous CTP releases, listeners were unusable because they did not persist when…
This blog post comes from a very interesting discussion with one of my friends about the read-only capabilities of secondary replicas in the context of distributed availability groups. Initially, distributed availability groups are designed to address D/R scenarios and some migration scenario types as well. I already discussed about of one possible migration scenario here. However, we may also take advantage of using secondary replicas as read-only in Reporting Scenarios (obviously after making an…
How to migrate an environment that includes availability groups from one Windows Failover Cluster to another one? This scenario is definitely uncommon and requires a good preparation. How to achieve this task depends mainly of your context. Indeed, we may use a plenty of scenarios according to the architecture in-place as well as the customer constraints in terms of maximum downtime allowed for example. Among all possible scenarios, there is a process called “cross-cluster migration…
In this blog let’s talk about SQL Server and availability groups in standard Edition. I had some customer questions about existing limitations and this is a good opportunity to review them on this blog post. First of all, let’s say that Microsoft has already documented the restrictions here and from my side, I already written a blog post on the subject at the time of the CTP version of SQL Server 2016. In the meantime,…
As promised in my latest blog, I will talk about improvements in the detection of the availability group replication latency. As a reminder, replication latency between high available replicas may be monitored in different manners so far. Firstly, in my experience I had the opportunity to use different performance counters as “Database Replica:Transaction Delay” and “Databas Replica:Mirrored Write Transactions/sec” in order to get a good picture of the latency over the time after implementing an…
As you certainly know, the SQL Server 2014 SP2 has been released by Microsoft with some interesting improvements that concern SQL Server AlwaysOn and availability groups feature. In fact, all of them are also included into SQL Server 2012 SP3 and SQL Server 2016 (update 24.09.2016: not yet released with SQL Server 2016). Among all fixes and improvements that concern AlwaysOn, I would like to focus on those described in the Microsoft KB3173156 and KB3112363. But in this…
As promised in my prior blog post, we’ll talk now about direct seeding feature and backups. We found an interesting LIMIT_CONCURRENT_BACKUPS value for current state of the hadr_automatic_seeding_state_transition event. As said last time, this value is meaningful by itself. So let’s try to perform a backup while the database is concerned by an active seeding session. In fact, my backup was blocked (as I supposed) and here what I found from the sys.dm_tran_locks and sys.dm_exec_requests…
This blog post follows the previous one about new direct seeding feature shipped with SQL Server 2016. As a reminder, I had some doubts about using direct seeding with large databases because log stream is not compressed by default but I forgot the performance improvements described into the Microsoft BOL. I also remembered to talk about it a couple of months ago in this blog post. So let’s try to combine all the things with the…
