Artifact download errors are not handled correctly

[adamgibbins]

Oct 30 13:05:36 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:13:in `info' resolving TIM-1.0.14519.jar
Oct 30 13:05:36 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:13:in `info' downloading artifact TIM-1.0.14519.jar from productstore.net.local
Oct 30 13:08:58 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:13:in `info' presentfalseapplicationTIMgroupblueparticipatingfalseversionhealth
Oct 30 13:10:36 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:21:in `error' AN ERROR OCCURED: #<Net::SCP::Error: SCP did not finish successfully (1)>
Oct 30 13:12:22 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:13:in `info' presentfalseapplicationTIMgroupblueparticipatingfalseversionhealth

In this instance for some reason SCP terminated mid-download, resulting in a corrupt artifact being dropped on the box. The app then failed to start, as expected. Future runs of Orc/deployapp fail as it retrieves the artifact from the cache, which is corrupt.

This then requires human intervention to remove the corrupt artifact.

Ideally deployapp should error when SCP terminates unexpectedly, rather than continuing until the app fails to start. Additionally it should not require a human to login to the box to rm the jar file in order to be able to successfully run Orc again, I'm not sure what it should do instead.

[tomwhoiscontrary]

lol

On Thu, 30 Oct 2014, Adam Gibbins wrote:

Oct 30 13:05:36 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:13:in `info' resolving TIM-1.0.14519.jar
Oct 30 13:05:36 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:13:in `info' downloading artifact TIM-1.0.14519.jar from productstore.net.local
Oct 30 13:08:58 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:13:in `info' presentfalseapplicationTIMgroupblueparticipatingfalseversionhealth
Oct 30 13:10:36 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:21:in `error' AN ERROR OCCURED: #<Net::SCP::Error: SCP did not finish successfully (1)>
Oct 30 13:12:22 latest-timcyclic-002 mcollectived[11446]: composite_logger.rb:13:in `info' presentfalseapplicationTIMgroupblueparticipatingfalseversionhealth

In this instance for some reason SCP terminated mid-download, resulting in a corrupt artifact being dropped on the box. The app then failed to start, as expected. Future runs of Orc/deployapp fail as it retrieves the artifact from the cache, which is corrupt.

This then requires human intervention to remove the corrupt artifact.

Ideally deployapp should error when SCP terminates unexpectedly, rather than continuing until the app fails to start. Additionally it should not require a human to login to the box to rm the jar file in order to be able to successfully run Orc again, I'm not sure what it should do instead.

---

Reply to this email directly or view it on GitHub:
#1

Information is not knowledge. -- Albert Einstein