Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

DB2 ibm_db gem support #531

Open
wants to merge 38 commits into
base: master
Choose a base branch
from
Open

DB2 ibm_db gem support #531

wants to merge 38 commits into from

Conversation

calh
Copy link

@calh calh commented Jul 18, 2019

This adds optional support for the ibm_db gem. If the ibm_db gem is not loaded in the bundle, octopus should behave the same as before. If the ibm_db gem is loaded, it enables some monkey patches in the ibm_db gem to work with Rails 4.2 and Rails 5.0.

This includes the Docker dev system commits in #526. Much of this PR is supporting code for the test suite and build system.

calh added 30 commits June 21, 2019 15:38
Running cucumber gives the error:

cannot load such file -- test/unit (LoadError)
These weren't supposed to make it into this PR
This is failing with the stacktrace:

Caused by:
LoadError: cannot load such file -- arel/visitors/bind_visitor
/usr/src/app/.bundle/ruby/2.5.0/gems/activesupport-5.2.3/lib/active_support/dependencies.rb:291:in `require'
/usr/src/app/.bundle/ruby/2.5.0/gems/activesupport-5.2.3/lib/active_support/dependencies.rb:291:in `block in require'
/usr/src/app/.bundle/ruby/2.5.0/gems/activesupport-5.2.3/lib/active_support/dependencies.rb:257:in `load_dependency'
/usr/src/app/.bundle/ruby/2.5.0/gems/activesupport-5.2.3/lib/active_support/dependencies.rb:291:in `require'
/usr/src/app/.bundle/ruby/2.5.0/gems/ibm_db-4.0.0/lib/active_record/connection_adapters/ibm_db_adapter.rb:13:in `<top (required)>'
/usr/src/app/.bundle/ruby/2.5.0/gems/activesupport-5.2.3/lib/active_support/dependencies.rb:291:in `require'
/usr/src/app/.bundle/ruby/2.5.0/gems/activesupport-5.2.3/lib/active_support/dependencies.rb:291:in `block in require'
/usr/src/app/.bundle/ruby/2.5.0/gems/activesupport-5.2.3/lib/active_support/dependencies.rb:257:in `load_dependency'
/usr/src/app/.bundle/ruby/2.5.0/gems/activesupport-5.2.3/lib/active_support/dependencies.rb:291:in `require'
/usr/src/app/lib/octopus/proxy_config.rb:245:in `initialize_adapter'
The gem install does this, but the download is slow.  Watching
the wget progress is more satisfying.
Error message is:

ActiveRecord::StatementInvalid: NoMethodError: undefined method `[]' for
nil:NilClass:

  from octopus/lib/octopus/abstract_adapter.rb:23:in `octopus_shard'
  from octopus/lib/octopus/abstract_adapter.rb:12:in `instrument'
  from vendor/bundle/ruby/2.3.0/gems/activerecord-4.2.11.1/lib/active_record/connection_adapters/abstract_adapter.rb:478:in `log'

This has shown to surface in the ibm_db and sqlserver adapters.
This allows the InstrumenterDecorator to properly log the shard name
Running tests individually works, but running the whole
migration_spec.rb file will recreate the issue _almost_
every time, in a random spec.

Here's the full error and stack trace:

Octopus::Migration should send the query to the correct shard
     Failure/Error:
       delegate :adapter_name, :add_transaction_record, :case_sensitive_modifier,
         :type_cast, :to_sql, :quote, :quote_column_name, :quote_table_name,
         :quote_table_name_for_assignment, :supports_migrations?, :table_alias_for,
         :table_exists?, :in_clause_length, :supports_ddl_transactions?,
         :sanitize_limit, :prefetch_primary_key?, :current_database,
         :combine_bind_parameters, :empty_insert_statement_value, :assume_migrated_upto_version,
         :schema_cache, :substitute_at, :internal_string_options_for_primary_key, :lookup_cast_type_from_column,
         :supports_advisory_locks?, :get_advisory_lock, :initialize_internal_metadata_table,
         :release_advisory_lock, :prepare_binds_for_database, :cacheable_query, :column_name_for_operation,
         :prepared_statements, :transaction_state, :create_table, to: :select_connection

     RuntimeError:
       An unexpected error occurred during retrieval of table metadata: uncaught throw :"Fetch Failure: <error message could not be retrieved>"
     # /usr/src/ruby-ibmdb/IBM_DB_Adapter/ibm_db/lib/active_record/connection_adapters/ibm_db_adapter.rb:1858:in `rescue in tables'
     # /usr/src/ruby-ibmdb/IBM_DB_Adapter/ibm_db/lib/active_record/connection_adapters/ibm_db_adapter.rb:1845:in `tables'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activerecord-5.0.7.2/lib/active_record/connection_adapters/abstract/schema_statements.rb:55:in `table_exists?'
     # ./lib/octopus/proxy.rb:25:in `table_exists?'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activerecord-5.0.7.2/lib/active_record/schema_migration.rb:20:in `block in table_exists?'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activesupport-5.0.7.2/lib/active_support/deprecation/reporting.rb:36:in `silence'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activesupport-5.0.7.2/lib/active_support/deprecation/instance_delegator.rb:20:in `silence'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activerecord-5.0.7.2/lib/active_record/schema_migration.rb:20:in `table_exists?'
     # /usr/src/ruby-ibmdb/IBM_DB_Adapter/ibm_db/lib/active_record/connection_adapters/ibm_db_adapter.rb:48:in `create_table'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activerecord-5.0.7.2/lib/active_record/connection_adapters/abstract/schema_statements.rb:1008:in `initialize_schema_migrations_table'
     # ./lib/octopus/proxy.rb:200:in `initialize_schema_migrations_table'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activerecord-5.0.7.2/lib/active_record/migration.rb:1127:in `initialize'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activerecord-5.0.7.2/lib/active_record/migration.rb:1017:in `new'
     # ./gemfiles/.bundle/ruby/2.5.0/gems/activerecord-5.0.7.2/lib/active_record/migration.rb:1017:in `run'
     # ./lib/octopus/migration.rb:187:in `block in run_with_octopus'
     # ./lib/octopus/proxy.rb:86:in `block (2 levels) in run_queries_on_shard'
     # ./lib/octopus/proxy.rb:345:in `using_shard'
     # ./lib/octopus/proxy.rb:85:in `block in run_queries_on_shard'
     # ./lib/octopus/proxy.rb:329:in `keeping_connection_proxy'
     # ./lib/octopus/proxy.rb:84:in `run_queries_on_shard'
     # ./lib/octopus/proxy.rb:93:in `block in send_queries_to_multiple_shards'
     # ./lib/octopus/proxy.rb:92:in `map'
     # ./lib/octopus/proxy.rb:92:in `send_queries_to_multiple_shards'
     # ./lib/octopus/migration.rb:186:in `run_with_octopus'
     # ./spec/support/octopus_helper.rb:46:in `migrate_to_version'
     # ./spec/support/octopus_helper.rb:37:in `migrating_to_version'
     # ./spec/octopus/migration_spec.rb:86:in `block (2 levels) in <top (required)>'
     # ------------------
     # --- Caused by: ---
     # UncaughtThrowError:
     #   uncaught throw :"\xA3\xC9\x8E+H\xEA\xC9\vX\x82g\x8B\xAAU\x00\x00\bY\r\x8C\xAAU\x00\x00\x1A\xD0\xC1\xE2\xCF\xE9\x1Dd0\xB3m\x8B\xAAU\x00\x00\x80T\r\x8C\xAAU\x00\x00\x01\x99\xD1o,"
     #   /usr/src/ruby-ibmdb/IBM_DB_Adapter/ibm_db/lib/active_record/connection_adapters/ibm_db_adapter.rb:1845:in `fetch_assoc'
calh added 8 commits July 18, 2019 11:35
This keeps the test suite going for most users that won't
be using the ibm_db gem
ibmdb/ruby-ibmdb#89

Until that issue is officially resolved by IBM, I'm using my local
fork for testing Octopus.
This is an oddball bug, deep in the DB2 C library.  Disconnecting
from the database during the Passenger fork process causes it to
hang indefinitely
Since we don't disconnect from DB2, this spec fails.
@thiagopradi
Copy link
Owner

@calh - any updates on this PR? would love to merge this changes.

@calh
Copy link
Author

calh commented Sep 25, 2019

Hi @thiagopradi! I think this PR should be good to go. I'm actually using it right now in production. The only caveat, is that it requires some changes to the ibm_db gem. https://github.com/ibmdb/ruby-ibmdb/pulls/calh

IBM will most likely pull these in, but their release cycle is pretty slow. The good news, is that the ibm_db gem is set up to be completely optional in this PR. So, people that are not using DB2 will never see my code anyway.

I'll leave it up to you if you want to wait for a new ibm_db version, or merge this in.

Thanks for you time!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants