SQL Server 2008 R2 shutdown / startup procedure

  • I have a SQL Server 2008 R2 2-node cluster running Windows Enterprise 2008 R2 and SQL Server standard Edition 2008 R2 using iSCSI attached NetApp storage. When either node is rebooted, a failover occurs normally and everything works accordingly. We recently needed to do some UPS upgrades on our server room and we took both nodes in the cluster down. Upon startup of node1, I noticed the cluster wouldn't come online and that in DISKPART, the iSCSI luns were set to RESERVED. I launched the Microsoft iSCSI initiator, and I was able to see that the targets were connected and the volumes were mapped. Because the WITNESS disk is one of those LUNs, the cluster wouldn't come online. I started Node2 and went into the iSCSI initiator and did a refresh on the target, and the LUNs then came online. I am not sure if it was the startup of Node2 or the refresh of the iSCSI initiator target that caused the disks to become available. So my question, is this normal behavior for having 2 nodes in a cluster shutdown down?

  • can you post results of the following when run from a command prompt on one of the nodes

    cluster /quorum

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • I have seen similar situations, particulary on my sandbox clusters.

    There are a few ways to safely shut down a cluster, and a few that can cause problems.

    Safe methods:

    a) Shut down each node one by one. Wait until all passive nodes have completed shutdown before shutting down the active node. On restart, start the previously active node first.

    b) Shut down SQL by putting the cluster offline (CLUSTER ClusterName RESOURCE GroupName /OFF) before you shutdown any nodes. After restart put the cluster back online, which will restart SQL.

    Unsafe methods:

    a) Stop SQL Server with NET STOP or similar. Almost always causes a problem.

    b) Shut down all nodes at the same time. This can cause a problem if some objects complete their failover before the active node shuts down, leaving other objects still on the active node. However, on a power failuer everything gets killed at once so this should not give a problem at restart.

    My recommendation is to build a sandbox cluster, then run some scenarios that could break it, and when it is broken work out how to fix it. Unless you have destroyed some data, just about all broken clusters are can be fixed, even if you have to eject nodes and add them back in again.

    Original author: https://github.com/SQL-FineBuild/Common/wiki/ 1-click install and best practice configuration of SQL Server 2019, 2017 2016, 2014, 2012, 2008 R2, 2008 and 2005.

    When I give food to the poor they call me a saint. When I ask why they are poor they call me a communist - Archbishop Hélder Câmara

Viewing 3 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic. Login to reply