January 18, 2018 at 11:58 am
We have some electrical work scheduled soon the will involve completely shutting down all power to a data center for a few hours. The work is between the UPS/generator system and the servers so the servers will all be dark during the work.
This includes several SQL Clusters (SQL 2008 through 2016 hosted on Windows 2008 RS and 2012 R2) where both nodes will be dark. What is the best practice for gracefully shutting down BOTH nodes of a cluster for the same time?
I'm looking at simply pausing the passive nodes to ensure everything comes back up on the same node, then shutting down the OS's, then resuming the paused nodes after startup.
Is it also necessary to manually shut down the SQL Instances before shutting down the machines, or will SQL Clustered instances gracefully and safely shut down before the OS shuts off?
January 18, 2018 at 12:50 pm
When we did a similar thing, we simply stopped the connectivity of the applications to the databases first, then went into FCM (Failover Cluster Manager) and executed a "Stop Role" for each of the instances in the cluster. Following this, we brought each node down (powered down) in the cluster. Later, when the nodes were powered back up, the instances (roles) automatically came online and I even think they did not need re-balancing.
January 19, 2018 at 5:33 am
dan-572483 - Thursday, January 18, 2018 11:58 AMWe have some electrical work scheduled soon the will involve completely shutting down all power to a data center for a few hours. The work is between the UPS/generator system and the servers so the servers will all be dark during the work.
This includes several SQL Clusters (SQL 2008 through 2016 hosted on Windows 2008 RS and 2012 R2) where both nodes will be dark. What is the best practice for gracefully shutting down BOTH nodes of a cluster for the same time?
I'm looking at simply pausing the passive nodes to ensure everything comes back up on the same node, then shutting down the OS's, then resuming the paused nodes after startup.
Is it also necessary to manually shut down the SQL Instances before shutting down the machines, or will SQL Clustered instances gracefully and safely shut down before the OS shuts off?
Yes clustered instances shutdown gracefully.
As previously stated above stopping the roles will take them offline, you will have to online the roles when the servers are back online.
There shouldn't be any need to pause the nodes
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
January 19, 2018 at 10:59 am
Some of clusters have multiple instances\roles, so it seems faster to pause & resume the passive nodes instead, and since this is every machine in that datacenter we'll have a lot of stuff to do.
Is there any downside to pausing\resuming the passives as opposed to stopping/restarting the roles?
January 22, 2018 at 11:02 am
pausing a node will drain roles and remove it temporarily from the cluster config, not exactly what you want to achieve.
Stop the roles and then restart them when the cluster is back online, you can even use simple PoSh scripts to do this
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply