Chef-metal and chef search


#1

I’ve been using chef-metal for a few days, and I’ve noticed that after the
initial instance creation, and chef-client run (initiated by chef-metal),
that, in order for the new node to be available in chef search results, I
have to manually log in to the instance and run chef-client a second
time. Waiting a few minutes isn’t a substitute. Why is this required?

Doug


#2

The Chef server takes some time before it indexes a node–it’s not so much
that you need a second run, as that you need your run to wait long enough
for solr to index the node. If you are having this problem with local
mode, that’s a different issue–chef-client -z should have nodes show up in
search results instantly.

On Wed, Oct 22, 2014 at 8:10 AM, Douglas Garstang doug.garstang@gmail.com
wrote:

I’ve been using chef-metal for a few days, and I’ve noticed that after the
initial instance creation, and chef-client run (initiated by chef-metal),
that, in order for the new node to be available in chef search results, I
have to manually log in to the instance and run chef-client a second
time. Waiting a few minutes isn’t a substitute. Why is this required?

Doug


#3

Is it possible that the first run doesn’t complete successfully?


Daniel DeLeo

On Wednesday, October 22, 2014 at 8:32 AM, John Keiser wrote:

The Chef server takes some time before it indexes a node–it’s not so much that you need a second run, as that you need your run to wait long enough for solr to index the node. If you are having this problem with local mode, that’s a different issue–chef-client -z should have nodes show up in search results instantly.

On Wed, Oct 22, 2014 at 8:10 AM, Douglas Garstang <doug.garstang@gmail.com (mailto:doug.garstang@gmail.com)> wrote:

I’ve been using chef-metal for a few days, and I’ve noticed that after the initial instance creation, and chef-client run (initiated by chef-metal), that, in order for the new node to be available in chef search results, I have to manually log in to the instance and run chef-client a second time. Waiting a few minutes isn’t a substitute. Why is this required?

Doug


#4

Hm.

I just tried to reproduce this and the nodes appeared within a few minutes
after their run. Pretty sure this wasn’t the behaviour I saw last night.
How’s the reliability of chef search? Are there known issues? I also had a
situation a few days ago where nodes weren’t appearing, uploads to my chef
server were resulting in an internal server error, and a reboot of the chef
server got rid of the errors and allowed searches to work correctly.

Doug.

On Wed, Oct 22, 2014 at 8:34 AM, Daniel DeLeo dan@kallistec.com wrote:

Is it possible that the first run doesn’t complete successfully?


Daniel DeLeo

On Wednesday, October 22, 2014 at 8:32 AM, John Keiser wrote:

The Chef server takes some time before it indexes a node–it’s not so
much that you need a second run, as that you need your run to wait long
enough for solr to index the node. If you are having this problem with
local mode, that’s a different issue–chef-client -z should have nodes show
up in search results instantly.

On Wed, Oct 22, 2014 at 8:10 AM, Douglas Garstang <
doug.garstang@gmail.com (mailto:doug.garstang@gmail.com)> wrote:

I’ve been using chef-metal for a few days, and I’ve noticed that after
the initial instance creation, and chef-client run (initiated by
chef-metal), that, in order for the new node to be available in chef search
results, I have to manually log in to the instance and run chef-client a
second time. Waiting a few minutes isn’t a substitute. Why is this
required?

Doug


Regards,

Douglas Garstang
http://www.linkedin.com/in/garstang
Email: doug.garstang@gmail.com
Cell: +1-805-340-5627