[jira] [Created] (DTACLOUD-99) deltacloudd is sometimes unresponsive to client queries

["Jaime Frey (Created) (JIRA)" <ji...@apache.org>]

deltacloudd is sometimes unresponsive to client queries
-------------------------------------------------------

                 Key: DTACLOUD-99
                 URL: https://issues.apache.org/jira/browse/DTACLOUD-99
             Project: DeltaCloud
          Issue Type: Bug
          Components: Server
         Environment: x86-64 Debian 6, x86-64 RHEL 6.1, ruby 1.9.2-p290
            Reporter: Jaime Frey
            Assignee: David Lutterkort
            Priority: Minor


When we test our deltacloud client code against the deltacloudd server with the mock driver, the server is unresponsive to client queries on a couple of our machines. The problem machines are running x86-64 Debian 6 and x86-64 RHEL 6.1. Other machines with these same OS's work just fine. In our test framework, we build ruby 1.9.2-p290 and install the necessary gems ourselves. Deltacloudd can be jolted back into normal operation by attaching to the child process running the thin gem with strace. As long as the process is being straced, it responds normally to client queries. As soon as strace detaches, deltacloudd returns to its unresponsive state.

Attaching to the child thin process with gdb doesn't change the server's behavior. But it does show that the process is spending all of it's time in EventMachine_t::Run (em.cpp:432), repeatedly calling EventMachine_t::_RunEpollOnce (em.cpp:473), which calls epoll_wait(). I haven't dug into that any further yet.

I don't know if this is a bug in ruby, one of the gems, or deltacloudd. I not know ruby and don't have the time to learn in order to track down this bug. We can provide access to the affected machines if someone is interested in investigating.

To reproduce on an affected machine:
* Build ruby-1.9.2-p290 and add to PATH
* Install gems:
  % gem install --no-rdoc --no-ri amazon-ec2 compass eventmachine haml json nokogiri rack rack-accept rack-test rake rcov rerun right_aws sinatra thin net-ssh
* Start deltacloudd:
  % deltacloud/server/bin/deltacloudd -i mock -p 3001
* Start client query:
  % wget --http-user=mockuser --http-passwd=mockpassword http://localhost:3001/api/instances

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira