Layout of deploy_revision


#1

Hi,

I’m confused about the directory structure that should be created by a
deploy_revision.

After perusing the Capistrano docs, I expected:

/my_app/releases/
/my_app/current - symlink to /my_app/releases/

instead I’m seeing

/my_app/releases/
/my_app/current/ - symlink to the matching revision under releases
/my_app/current/current - symlink to the latest revision under releases

This seems odd - but from the docs I can’t figure out if that’s what I
should be expecting.

Any pointers to how this should work?

Malcolm


#2

On Oct 4, 2011, at 4:34 AM, Malcolm Box boxm+chef@livetalkback.com wrote:

Hi,

I’m confused about the directory structure that should be created by a deploy_revision.

After perusing the Capistrano docs, I expected:

/my_app/releases/
/my_app/current - symlink to /my_app/releases/

instead I’m seeing

/my_app/releases/
/my_app/current/ - symlink to the matching revision under releases
/my_app/current/current - symlink to the latest revision under releases

This seems odd - but from the docs I can’t figure out if that’s what I should be expecting.

Any pointers to how this should work?

Malcolm

What you are seeing is not the expected behavior. It’s hard to say why your problem is occurring. Perhaps run ‘chef-client -l debug’ for more detail?


#3

On 4 October 2011 14:21, Three Tee threetee@gmail.com wrote:

On Oct 4, 2011, at 4:34 AM, Malcolm Box boxm+chef@livetalkback.com
wrote:

instead I’m seeing

/my_app/releases/
/my_app/current/ - symlink to the matching revision under
releases
/my_app/current/current - symlink to the latest revision under releases

What you are seeing is not the expected behavior. It’s hard to say why your
problem is occurring. Perhaps run ‘chef-client -l debug’ for more detail?

Thanks! Now I’ve re-run it again and am getting what I expected - current
linking directly to a release. I’d swear it was working differently
yesterday. Very odd.

Malcolm


#4

On Oct 4, 2011, at 2:46 PM, Malcolm Box wrote:

Thanks! Now I’ve re-run it again and am getting what I expected - current linking directly to a release. I’d swear it was working differently yesterday. Very odd.

You might try running it on a “fresh” server, i.e., one where you had not run into this problem before.

One thing I’ve discovered is that sometimes Chef will run into what it thinks are problems and bomb out, but in fact what it wanted to have done was actually accomplished, so that all you need to do is run it again. In particular, I’ve seen this behaviour with Chef and yum, which sometimes exits with non-zero codes even if it did what you asked it to do.

I don’t know if that is the intended behaviour of Chef, but it is what I have observed.


Brad Knowles bknowles@ihiji.com
SAGE Level IV, Chef Level 0.0.1


#5

It is intended - in the face of failure, we abort, and if it’s
ephemeral, idempotency and convergence will get you where you want to
go.

Adam

On Tue, Oct 4, 2011 at 3:21 PM, Brad Knowles bknowles@ihiji.com wrote:

On Oct 4, 2011, at 2:46 PM, Malcolm Box wrote:

Thanks! Now I’ve re-run it again and am getting what I expected - current linking directly to a release. I’d swear it was working differently yesterday. Very odd.

You might try running it on a “fresh” server, i.e., one where you had not run into this problem before.

One thing I’ve discovered is that sometimes Chef will run into what it thinks are problems and bomb out, but in fact what it wanted to have done was actually accomplished, so that all you need to do is run it again. In particular, I’ve seen this behaviour with Chef and yum, which sometimes exits with non-zero codes even if it did what you asked it to do.

I don’t know if that is the intended behaviour of Chef, but it is what I have observed.


Brad Knowles bknowles@ihiji.com
SAGE Level IV, Chef Level 0.0.1


Opscode, Inc.
Adam Jacob, Chief Product Officer
T: (206) 619-7151 E: adam@opscode.com


#6

On Wed, Oct 5, 2011 at 12:38 AM, Adam Jacob adam@opscode.com wrote:

It is intended - in the face of failure, we abort, and if it’s
ephemeral, idempotency and convergence will get you where you want to
go.

Well, that’s probably true on master [*], but on 0.10.4 there are
plenty of things that can go wrong.


#7

DO WANT

/me flaps hand open and closed

On 6 October 2011 21:00, Andrea Campi andrea.campi@zephirworks.com wrote:

On Wed, Oct 5, 2011 at 12:38 AM, Adam Jacob adam@opscode.com wrote:

It is intended - in the face of failure, we abort, and if it’s
ephemeral, idempotency and convergence will get you where you want to
go.

Well, that’s probably true on master [*], but on 0.10.4 there are
plenty of things that can go wrong.