Month / May 2012

by Arvind Shyamsundar

“Failed to join domain” trying to add Windows 2008 R2 Core to domain

Today I was using SCONFIG (if you don’t know what SCONFIG is, check out these posts to add a Windows 2008 R2 core server to an existing domain. I had set the IP address and ensured that the domain name and credentials are correct, but I kept getting ‘failed to join domain’.

I tried pre-creating the computer objects in the AD but to no avail. Finally after some troubleshooting it turned out to be a most basic issue: the primary DNS IP was not set correctly. After setting that (again with SCONFIG) I was able to add the Windows core box into the domain.

Advertisement
by Arvind Shyamsundar

‘No Disks suitable for cluster disks’: one more possible cause

Today in the lab I was working on a test 3-node cluster. I was using Hyper-V for this. We normally use the iSCSI Software Target for testing, so I had set that up on a central VM and then published the target to all the VMs.

The disks were seen in the disk management of Node1, and I brought them online, partitioned and formatted them. However when I used Cluster Management to add them to the Storage group, it complained:

‘No Disks suitable for cluster disks were found, For diagnostic information about disks available run the Validate config wizard’

I did run the cluster validation (storage only) and got no further inputs. Shortly after some troubleshooting I realized I had forgotten to configure the volumes at the iSCSI initiator on each node. The easiest thing to do for this is to click the ‘Auto-configure’ button:

image

Due disclaimer: This is just ONE, but not the ONLY reason for the above error message! This is somewhat specific to the iSCSI initiator, so only check it if you are using iSCSI volumes in your cluster and if you get the above message.

by Arvind Shyamsundar

Opinion Poll: Are PRINT statements considered harmful?

Today during a discussion a point came up around the role of PRINT statements in production code. While most data access today is routed through a data access layer (typically .NET or JDBC) and is focussed on consuming result sets (or executing U/I/D nonquery stuff) we were wondering on what you use PRINT statements for. In the long past, I would have said that the print statement is probably the best way for debugging, but in today’s world with easy access to the T-SQL debugger, a developer is probably much better off without PRINT.

With SqlClient, we consume the PRINT events using the InfoMessage event, but in practice I wonder how many of you actually rely on that. Please comment on this post to share your experiences with PRINT and if you think it is good / bad / evil in today’s world 🙂