custodian.git - A distributed protocol tester.

Age	Commit message (Collapse)	Author
2016-04-22	More rubocop fixups.	Steve Kemp
	These are again mostly based around whitespace-changes.
2016-01-18	Updated the queue-handling.	Steve Kemp
	We now use a zset to store our pending tests. This means that jobs are only in the queue once - no duplicates are allowed. This closes #12428.
2015-12-18	Updated to use the right form of counting for the set.	Steve Kemp

2015-12-18	Fixed the name of the string.	Steve Kemp

2015-12-18	Updated to revert to a set with no ordering.	Steve Kemp
	This is more reliable, albeit potentially racy and with the failure case that a job might be readded twice.
2015-12-18	Return values using a reverse-score-range.	Steve Kemp
	This prevents starvation, by ensuring that we pull tests out in a FIFO fashion - by virtue of the timestamp.
2015-12-18	Removed references and support for beanstalkd.	Steve Kemp
	The beanstalkd queue used to be used in the past, and we later added support for Redis via a simple abstraction layer. But now we've no longer tested and used beanstalkd for over a year, and the client-libraries are no longer available as native Debian packages. With that in mind we've excised the code, although left the abstraction-class in-place.
2015-12-18	Removed the diagnostic output of the test-scores	Steve Kemp

2015-12-18	Use a sorted set for tests in our queue.	Steve Kemp
	This ensures that all tests always run, and we have an ordering.
2015-12-17	Treat our Redis queue as a set.	Steve Kemp
	This means that tests will only ever be enqueued once, regardless of how many times they are parsed. In the past we could have a configuration file that read: test1 .. test2 .. test3 .. Parsing/adding this file would result in a queue looking like so: test1 .. test2 .. test3 .. test1 .. test2 .. test3 .. test1 .. test2 .. test3 .. Now the queue will ALWAYS look like this: test1 .. test2 .. test3 .. In the normal course of events this won't matter, as teh processing loop will look like so: * Add new jobs every minute. * Worker runs the jobs. In the case of a failing job though the test might take 2.5 minutes and that will cause the queue to backup. (2.5 minutes because a test is repeated 5 times before a fail is announced, and the timeout is 30 seconds. These values can and should be tweaked.) With the new method even if the queue is slowly draining the queue will never grow to containu hundreds of events it will just be "topped up" not "overflowing". Thanks to James Hannah for the suggestion, and James Lawrie for the patience.
2015-04-16	Mark variables as unused.	Steve Kemp
	We've now prefixed our transient/unused variables with "_" to stop any future errors about variables not being used.
2015-03-09	Removed trailing whitespace from the code	Steve Kemp

2015-03-09	Remove spaces inside blocks.	Steve Kemp

2015-03-09	Whitespace fixups.	Steve Kemp
	These were all identified and suggested by rubocop.
2015-03-09	Use def with parentheses when there are parameters.	Steve Kemp

2015-03-09	Removed spaces inside parenthesis.	Steve Kemp

2015-03-09	Prefer single-quotes when you don't need interpolation.	Steve Kemp
	So "foo" is less good than 'foo'.
2015-03-09	Don't use parenthesis aroudn conditions in an if.	Steve Kemp

2015-03-09	Do not use parentheses for method calls with no arguments.	Steve Kemp
	This is neater. Flagged by rubocop
2015-03-09	Omit the parentheses in defs when the method doesn't accept any arguments.	Steve Kemp
	This is neater.
2015-03-09	Avoid arguments we're not using.	Steve Kemp
	Having methods take arguments which are ignored is a misleading thing, prefix with "_" to make that explicit, or remove.
2015-02-18	Prefer "while true" to "while 1"	Steve Kemp
	Also only sleep between fetches if we couldn't find a job because the queue was empty.
2015-02-18	Minor formatting and commenting update.	Steve Kemp

2015-02-18	Updated the way we wait for jobs.	Steve Kemp
	The new approach uses the redis gems timeout functionality and ensures we never return a null-job. Instead we timeout and repeat with a stalling-sleep in the way. This closes #9553.
2015-02-17	fix the redis queue not to melt the cpu when polling redis for a job	Patrick J Cherry

2015-02-05	Simplify code by removing configuration.	Steve Kemp
	We now set the queue-address via $QUEUE_ADDRESS, otherwise we default to localhost. This works for both redis & beanstalkd.
2015-02-05	Simplify the flushing of queues.	Steve Kemp
	Now that we've moved to using redis by default the handling of queue-flushing needs to change. We can simply get rid of the busy-wait and run a redis "del" operation. With that in mind we've moved the flushing logic to our queue abstraction layer, and simplified our queue-helper script.
2015-01-29	Don't output messages unless $DEBUG=1	Steve Kemp

2015-01-19	Updated to use our queue-abstraction.	Steve Kemp
	This means we can queue/dequeue to either Redis or Beanstalkd.
2015-01-13	Added queue-abstraction layer.	Steve Kemp
	This is not-yet used, but it will-be shortly. The intention is that we can seamlessly swap out the queue implemention in the near future so that we'll be able to use Redis.
2012-11-20	Removed	Steve Kemp

2012-11-20	New stub interface.	Steve Kemp