Permalink
Browse files

Revert "Use excon gem", DJ stacked up 4,000+ jobs since this caused S…

…SL errors

This reverts commit 968eedd.
  • Loading branch information...
1 parent 78ea644 commit 5d47b91b86ace036349cf8f1a2516418c354cad8 @qrush qrush committed Dec 16, 2011
View
@@ -4,7 +4,12 @@ gem 'rails', '~> 3.0.10'
gem 'airbrake'
gem 'clearance'
-gem 'excon'
+# Use a fork of excon to resolve the issue discussed in:
+# https://github.com/geemus/excon/pull/87
+#
+# It's should be safe to remove this line once this pull request is merged and
+# a new gem version of excon is released.
+gem 'excon', :git => 'git://github.com/evanphx/excon.git'
gem 'fog'
gem 'gchartrb', :require => 'google_chart'
gem 'gravtastic'
View
@@ -1,3 +1,9 @@
+GIT
+ remote: git://github.com/evanphx/excon.git
+ revision: 24ffc60420c09cecb46e1d7c177a8919627057a9
+ specs:
+ excon (0.7.12)
+
GEM
remote: https://rubygems.org/
specs:
@@ -79,21 +85,21 @@ GEM
diff-lcs (1.1.3)
erubis (2.6.6)
abstract (>= 1.0.0)
- excon (0.9.1)
factory_girl (2.3.2)
activesupport
factory_girl_rails (1.4.0)
factory_girl (~> 2.3.0)
railties (>= 3.0.0)
ffi (1.0.11)
- fog (0.7.2)
+ fog (1.1.1)
builder
- excon (>= 0.6.1)
- formatador (>= 0.1.3)
- json
+ excon (~> 0.7.4)
+ formatador (~> 0.2.0)
mime-types
+ multi_json (~> 1.0.3)
+ net-scp (~> 1.0.4)
net-ssh (>= 2.1.3)
- nokogiri (>= 1.4.4)
+ nokogiri (~> 1.5.0)
ruby-hmac
formatador (0.2.1)
gchartrb (0.8)
@@ -229,7 +235,7 @@ DEPENDENCIES
database_cleaner
delayed_job (~> 3.0.0.pre)
delayed_job_active_record
- excon
+ excon!
factory_girl_rails
fog
gchartrb
Binary file not shown.
Binary file not shown.
Binary file not shown.

8 comments on commit 5d47b91

@sferik
Member
sferik commented on 5d47b91 Dec 16, 2011

Strange. I thought the latest gem release included all of @evanphx's changes. Is there an easy way for me to keep an eye on the size of the DJ queue (without SSHing into the server)?

@bryckbost

Might be worth taking a look at https://github.com/ejschmitt/delayed_job_web

@geemus
geemus commented on 5d47b91 Dec 16, 2011

Evan pushed another fix having to due with OpenSSL + write_nonblock madness here:

excon/excon#95 (comment)

I have since merged and released 0.9.2 (with which you will hopefully not have this problem). Let me know if there is anything else I can do to help. Thanks!

@qrush
Member
qrush commented on 5d47b91 Dec 16, 2011

I'm fine with trying this, as long as it's verified on staging that the indexing and pushing a gem to S3/CF works.

That being said we need some kind of CI/canary check to make sure this doesnt happen anymore.

@sferik
Member
sferik commented on 5d47b91 Dec 17, 2011

Okay, I upgraded to 0.9.2 (82386ff) and pushed to staging.

I'll be testing it thoroughly on staging before pushing to production. If anyone wants to help testing, that would be welcome. :)

@sferik
Member
sferik commented on 5d47b91 Dec 17, 2011

Okay, I successfully pushed hola version 0.0.20 to staging: http://staging.rubygems.org/gems/hola/versions/0.0.20

I was able to install the gem from staging, however the download count did not increment. Any idea why that might be? Are those download counts cached?

@qrush
Member
qrush commented on 5d47b91 Dec 18, 2011

The download count bumped fine for me....strange. 😒

@sferik
Member
sferik commented on 5d47b91 Dec 19, 2011

Hmmm, I just pushed a new version and still can't get the download count to bump. What command are you using? I tried:

RUBYGEMS_HOST=https://staging.rubygems.org gem install hola
Please sign in to comment.