Jira (PUP-2738) Investigate FFI Memory Pressure / Deterministically Release FFI MemoryPointer

[Ethan Brown (JIRA)]

Title: Message Title

Ethan Brown created an issue   Puppet / PUP-2738 Investigate FFI Memory Pressure / Deterministically Release FFI MemoryPointer Issue Type: Task Assignee: Unassigned Created: 06/Jun/14 3:21 PM Priority: Major Reporter: Ethan Brown As of the merge for PUP-2657 , we've noticed some intermittent test failures (in both CI and locally) in semi-random locations. The merge commit is as follows: https://github.com/puppetlabs/puppet/commit/d38153bd712167e2777bdd4c63d37e2e8d1cdb39 The commit itself has been reviewed and it doesn't seem to have issues itself, but it's likely it's triggered FFI memory pressure that might have been building up to this point. Locally, failures include rspec ./spec/unit/util/storage_spec.rb:178 # Puppet::Util::Storage when storing to the state file should create the state file if it does not exist rspec ./spec/unit/util/storage_spec.rb:196 # Puppet::Util::Storage when storing to the state file should load() the same information that it store()s rspec ./spec/unit/util/yaml_spec.rb:16 # Puppet::Util::Yaml writes data formatted as YAML to disk rspec ./spec/unit/util_spec.rb:438 # Puppet::Util#replace_file should copy 0555 permissions from the target file by default rspec ./spec/unit/util_spec.rb:438 # Puppet::Util#replace_file should copy 0660 permissions from the target file by default rspec ./spec/unit/util_spec.rb:438 # Puppet::Util#replace_file should copy 0770 permissions from the target file by default rspec ./spec/unit/util_spec.rb:465 # Puppet::Util#replace_file should use the default permissions if the source file doesn't exist rspec ./spec/unit/util_spec.rb:503 # Puppet::Util#replace_file should support string format permissions rspec ./spec/unit/util_spec.rb:503 # Puppet::Util#replace_file should support number format permissions rspec ./spec/unit/util_spec.rb:503 # Puppet::Util#replace_file should support symbolic format permissions Ideally we should be running windbg and monitoring heap allocations to very bad behavior. Another idea is to use the block form of the FFI MemoryPointer initializer, and ensure that memory is deterministically freed, rather than when the Ruby GC decides to free it up. Add Comment   This message was sent by Atlassian JIRA (v6.1.4#6159-sha1:44eaede) -- You received this message because you are subscribed to the Google Groups "Puppet Bugs" group. To unsubscribe from this group and stop receiving emails from it, send an email to puppet-bugs+unsubscr...@googlegroups.com. To post to this group, send email to puppet-bugs@googlegroups.com. Visit this group at http://groups.google.com/group/puppet-bugs. For more options, visit https://groups.google.com/d/optout.