| Author |
Topic |
|
kdubb777
Starting Member
24 Posts |
 Posted - 2008-02-11 : 16:46:54
|
| Does anyone know of any issues with creating only one group for all resources involved in a two node SQL failover cluster???So in this one and only group we have ClusterIP,ClusterName,Q:E:, MSDTC,SQL SERVER, SQL AGENT etc...I know it's not best practices. I have never seen a cluster setup this way but it is up and running!! heh!! LOL!!Kevin- |
|
|
eyechart
Master Smack Fu Yak Hacker
3575 Posts |
Posted - 2008-02-11 : 16:57:54
|
| are there separate IPs for the clustername, the DTC name and the SQL Server?-ec |
 |
|
|
Haywood
Posting Yak Master
221 Posts |
Posted - 2008-02-11 : 17:14:24
|
| It'll work, but it limits you in what you can do with the system and it's availability. You cannot move an individual resource without affecting the group as a whole and it's other individual resources... |
|
|
|
kdubb777
Starting Member
24 Posts |
Posted - 2008-02-11 : 17:52:06
|
| Yeah all the resources will have to shift from one node to the other atomically in this scenario. But since this cluster is only a two node failover cluster running one instance I guess it's not that bad. Taking the SQL server offline will take the cluster service offline though. |
|
|
|
kdubb777
Starting Member
24 Posts |
Posted - 2008-02-11 : 17:52:44
|
| There are separate IP's and names for the cluster name and IP and SQL name and IP |
|
|
|
Haywood
Posting Yak Master
221 Posts |
Posted - 2008-02-11 : 17:57:21
|
| If you can get ten to fifteen minutes of downtime, you can split them up into multiple groups and move the appropriate resource to the appropriate group...it's pretty easy to fix this "issue". :) |
|
|
|
kdubb777
Starting Member
24 Posts |
Posted - 2008-02-11 : 18:27:13
|
| Yeah but I don't think we can pyhsicaly split them because the server was only presented one 400 GB LUN. The sys admins then partitioned the LUN into 3 partitions so the cluster sees Q, L and G all as one physical disk. I don't think it is possible. |
|
|
|
Haywood
Posting Yak Master
221 Posts |
Posted - 2008-02-11 : 18:27:57
|
| Ah, yeah, you're hosed... |
|
|
|
kdubb777
Starting Member
24 Posts |
Posted - 2008-02-11 : 18:32:04
|
| But since it's only one instance and it's only a SQL cluster I guess it's not that bad. If the cluster service fails or the quorum drive fails SQL will be insacessible regardless of the grouping. |
|
|
|
rmiao
Master Smack Fu Yak Hacker
7266 Posts |
Posted - 2008-02-11 : 22:55:14
|
| Not that bad. In fact I set my clusters that way since I like all resources owned by single node, much easier for system maintenance. By the way, ssms will show dtc is down if it owned by different node than sql resources. |
|
|
|
kdubb777
Starting Member
24 Posts |
Posted - 2008-02-26 : 17:40:18
|
| Yeah but what about the quorum drive and best practices. Best practices are set out for a reason but they are by no means the law. Quorum data and SQL data being on the same physical disk I would think would come back to bite you at some point. |
|
|
|
rmiao
Master Smack Fu Yak Hacker
7266 Posts |
Posted - 2008-02-26 : 22:01:32
|
| You like to put quorum on 72gb disk alone? Sure you can, but I don't want to waste 71gb disk space. |
|
|
|
|
2 node SQL cluster all resources in one group