[chef] Re: Re: High Availability of Bookshelf in Chef 11


Chronological Thread 
  • From: Mark Mzyk < >
  • To:
  • Subject: [chef] Re: Re: High Availability of Bookshelf in Chef 11
  • Date: Tue, 10 Jun 2014 21:10:34 -0400

Bookshelf conforms the S3 api, so it's possible to use S3 itself instead of bookshelf.

See the bookshelf options here: http://docs.opscode.com/config_rb_chef_server_optional_settings.html#bookshelf

You'll want to set the bookshelf url to point to an S3 bucket and set the access and secret keys appropriately, then do a chef-server-ctl reconfigure so the settings are picked up and you should be good to go.

At that point you can even disable the bookshelf service if you want, as the server will pick up the settings and talk directly to S3, so bookshelf itself doesn't have to be running.

- Mark

" type="cite">
" photoname="tetsu soh" src="jpgJeuo9olkAS.jpg" name="compose-unknown-contact.jpg" height="25px" width="25px">
June 10, 2014 at 8:19 PM
Hi Adam,

Thank you!

One more question, is it possible to use S3 as a bookshelf service? 

Regards,

Tetsu

2014年6月11日水曜日、Adam Jacob< "> >さんは書きました:


--
Sent from Gmail Mobile
" photoname="Adam Jacob" src="jpgZWnw0XRDgd.jpg" name="postbox-contact.jpg" height="25px" width="25px">
June 10, 2014 at 7:53 PM
No, Bookshelf doesn't sync the data for you. That configuration option maps to the HA configuration of Enterprise Chef, and it essentially disables the service by default and hands control over its state to a supervisor. We use DRBD to replicate the block device between two nodes, and keepalived to transition a VIP between them.

Adam





--
Opscode, Inc.
Adam Jacob, Chief Dev Officer
T: (206) 619-7151 E: ">
" photoname="tetsu soh" src="jpgJeuo9olkAS.jpg" name="compose-unknown-contact.jpg" height="25px" width="25px">
June 10, 2014 at 11:54 AM
Hello,

According to document[1] of Chef 11 Server, there is an option in Bookshelf part

bookshelf['ha'] : 
Use to run the Chef server in a high availability topology. Default value: false.

What does this mean? Does the HA here mean we can setup a bookshelf cluster with multiple nodes running Bookshelf service and Bookshelf will sync data (replica) automatically for us?
If it is true, how can I specify nodes in the cluster?
If it is false, how can I set up a cluster of Bookshelf ?
Can I use other S3 service instead of Bookshelf ?

Thank you for your help.

Regards,

Tetsu

[1] http://docs.opscode.com/config_rb_chef_server_optional_settings.html



Archive powered by MHonArc 2.6.16.

§