Clusters?


#1

How does one deal with a “shared” cluster IP - ie one that can move in between a group of nodes?

I have a SQL Server Failover Cluster, and was going to install chef on the nodes in order to make them usable in recipes, but realized that I would have the node IPs, not the virtual ones…

How do we deal with this?

Thanks.

-Matt


#2

On Friday, June 24, 2011 at 8:31 AM, Matthew Drobnak wrote:

How does one deal with a “shared” cluster IP – ie one that can move in between a group of nodes?

I have a SQL Server Failover Cluster, and was going to install chef on the nodes in order to make them usable in recipes, but realized that I would have the node IPs, not the virtual ones…

How do we deal with this?

Thanks.

-Matt
We track these in data bags, which is a bit manual, but the IPs are long lived so it’s not much of a hassle. Then, for anything that we configure to talk to them, they access the VIP out of the data bag.


Dan DeLeo


#3

That’s what I did, but I added a new item in my data bag, and checked for it’s presence if the search yielded 0 results…

Is that what you are doing (ie modifying the recipes)?

Thanks.

-Matt

-----Original Message-----
From: Daniel DeLeo [mailto:ddeleo@kallistec.com] On Behalf Of Daniel DeLeo
Sent: Friday, June 24, 2011 11:36 AM
To: chef@lists.opscode.com
Subject: [chef] Re: Clusters?

On Friday, June 24, 2011 at 8:31 AM, Matthew Drobnak wrote:

How does one deal with a “shared” cluster IP – ie one that can move in between a group of nodes?

I have a SQL Server Failover Cluster, and was going to install chef on the nodes in order to make them usable in recipes, but realized that I would have the node IPs, not the virtual ones…

How do we deal with this?

Thanks.

-Matt
We track these in data bags, which is a bit manual, but the IPs are long lived so it’s not much of a hassle. Then, for anything that we configure to talk to them, they access the VIP out of the data bag.


Dan DeLeo


#4

On Friday, June 24, 2011 at 8:37 AM, Matthew Drobnak wrote:

That’s what I did, but I added a new item in my data bag, and checked for it’s presence if the search yielded 0 results…

Is that what you are doing (ie modifying the recipes)?

Thanks.

-Matt
Yeah, we have data_bags/environments/{prod,preprod, … }.json

We access them in the recipes with something like

env = data_bag_item(“environments”, node.chef_environment)

And the vips are like

env[“foo_app_vip”]

Anyway, if you’re using git and cookbooks from the cookbooks site, use the vendor branch stuff with knife cookbook install to install the cookbooks into your git repo, you can modify the cookbooks while still tracking the upstream.


Dan DeLeo