Gosh, I learn something new every day. I'll check out bundle_package. If
one uses that, it looks like it just work automagically, meaning Bundler
just understands that it needs to use the gem cache instead of
rubygems.org. That's nifty. Two questions for you:
1) Do I update the cache when I change/add a gem by running "bundle
install" and then "bundle package"? That's a little unclear on the page you
linked to.
2) Do I have to do anything to tell Cap to use the bundle cache, maybe set
:bundle_flags to include "--local"? The page says this:
By default, if you simply run bundle install after running bundle package,
Bundler will still connect to rubygems.org to check whether a
platform-specific gem exists for any of the gems in vendor/cache. This
behavior can be avoided by instead running bundle install --local. Note
that this requires you to have the correctly platformed version for all of
your gems already cached. The easiest way to achieve this is to run bundle
package on an identical machine and then check in those vendored gems.
Makes me think that little "--local" tidbit has to make an appearance
somewhere.
Thanks for the recommendation.
On Friday, April 4, 2014 11:53:13 AM UTC-4, Lee Hambley wrote:
>
> That sounds reasonable to me. I did introduce some new gems, so that might
>> have pushed the memory usage over the edge.
>>
>
> It happens, the weirder one is when the oomkiller kills bundler because
> you are *almost* out of memory, but you can only see the reason in some
> obscure non-standard log file.
>
> I'd love to keep using the free t1.micros, at least for testing. You
>> mention server-side bundling. Is there a way to "pre-bundle" stuff so that
>> the bundle install portion of the Cap deploy doesn't introduce this heavy
>> memory load (or take as long)? I saw some mention of being able to do this
>> ("locking" a bundle, or some such) in the past, but I gathered it's not
>> possible anymore. Or I just don't understand things well enough yet. That's
>> always a possibility.
>>
>
> Check out http://bundler.io/v1.5/bundle_package.html - it will pack all
> the gems into your Git repo, which will make it balloon a bit, but if your
> choice is "slightly slower clone/checkout" vs. "pay more", I know what I'd
> take for testing projects.
>
> Good luck,
>
> - Lee
>
>
>> O
>> n Friday, April 4, 2014 11:45:36 AM UTC-4, Lee Hambley wrote:
>>
>>> Probably you introduced new Gems, or some gems are updated, or the
>>> Rubygems index metadata just got too large. One can't really expect to
>>> server-side bundle on a t1.micro reliably.
>>>
>>> Lee Hambley
>>> --
>>> http://lee.hambley.name/
>>> +49 (0) 170 298 5667
>>>
>>>
>>> On 4 April 2014 17:44, Roy Miller <[email protected]> wrote:
>>>
>>>> Serves me right for not running my deploy for 10 days, but now when I
>>>> run it I get this:
>>>>
>>>> [15:30:41] INFO [61afb781] Running */usr/bin/env bundle install
>>>> --binstubs /var/www/wait-less/shared/bin --path
>>>> /var/www/wait-less/shared/bundle --without development --deployment* on
>>>> ec2-107-23-0-121.compute-1.amazonaws.com[15:30:41] DEBUG [61afb781]
>>>> Command: cd /var/www/wait-less/releases/20140404153040 && /usr/bin/env
>>>> bundle install --binstubs /var/www/wait-less/shared/bin --path
>>>> /var/www/wait-less/shared/bundle --without development --deployment
>>>>
>>>> [15:30:43] DEBUG [61afb781] Fetching source index from
>>>> https://rubygems.org/
>>>>
>>>> [15:32:21] DEBUG [61afb781] Retrying git cat-file -e
>>>> 3d1afeb52e974246ec5dbe23c12147d1116195d0 due to error (2/3): Errno::ENOMEM
>>>> Cannot allocate memory - git[15:32:21] DEBUG [61afb781] Retrying git
>>>> cat-file -e 3d1afeb52e974246ec5dbe23c12147d1116195d0 due to error (3/3):
>>>> Errno::ENOMEM Cannot allocate memory - git
>>>>
>>>>
>>>> Looks like Bundler is running out of memory for some reason. I'm
>>>> running on an AWS t1.micro instance, which could be the issue, but it
>>>> worked just dandy 10 days ago, with no changes at all to the box, only
>>>> code
>>>> changed.
>>>>
>>>> Any ideas about what's causing the issue, and how to fix it?
>>>>
>>>> Roy
>>>>
>>>> --
>>>> You received this message because you are subscribed to the Google
>>>> Groups "Capistrano" group.
>>>> To unsubscribe from this group and stop receiving emails from it, send
>>>> an email to [email protected].
>>>>
>>>> To view this discussion on the web, visit https://groups.google.com/d/
>>>> msgid/capistrano/7e3d55cf-7623-47f6-831d-6b9622ab7265%
>>>> 40googlegroups.com<https://groups.google.com/d/msgid/capistrano/7e3d55cf-7623-47f6-831d-6b9622ab7265%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>> .
>>>> For more options, visit https://groups.google.com/d/optout.
>>>>
>>>
>>> --
>> You received this message because you are subscribed to the Google Groups
>> "Capistrano" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected] <javascript:>.
>> To view this discussion on the web, visit
>> https://groups.google.com/d/msgid/capistrano/357dcc56-bb6b-491f-b519-87e4c108f102%40googlegroups.com<https://groups.google.com/d/msgid/capistrano/357dcc56-bb6b-491f-b519-87e4c108f102%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>
--
You received this message because you are subscribed to the Google Groups
"Capistrano" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web, visit
https://groups.google.com/d/msgid/capistrano/4d8b11f4-951e-4490-8ae1-0f881cc05c99%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
