(POOLER-191) Add checking for running instances that are not in active

This change adds detection of running instances that are in a running
queue, but have no data in a active queue for the same pool. When this
happens a machine will live forever, impacting the running count, and
preventing the machine from being killed. Without this change running
instances that are not marked as active will live forever.
This commit is contained in:
kirby@puppetlabs.com 2020-10-13 13:59:13 -07:00
parent 35104a75bb
commit 0ad069b958
2 changed files with 17 additions and 13 deletions

View file

@ -289,23 +289,19 @@ module Vmpooler
move_vm_queue(pool, vm, 'running', 'completed', redis, "reached end of TTL after #{ttl} hours")
throw :stop_checking
end
else
move_vm_queue(pool, vm, 'running', 'completed', redis, 'is listed as running, but has no checkouttime data. Removing from running')
end
if provider.vm_ready?(pool, vm)
throw :stop_checking
else
host = provider.get_vm(pool, vm)
throw :stop_checking if provider.vm_ready?(pool, vm)
throw :stop_checking if provider.get_vm(pool, vm)
if host
throw :stop_checking
else
move_vm_queue(pool, vm, 'running', 'completed', redis, 'is no longer in inventory, removing from running')
end
end
end
end
end
end
def move_vm_queue(pool, vm, queue_from, queue_to, redis, msg = nil)
redis.smove("vmpooler__#{queue_from}__#{pool}", "vmpooler__#{queue_to}__#{pool}", vm)

View file

@ -647,12 +647,20 @@ EOT
end
context 'valid host' do
it 'should not move VM if it has no checkout time' do
it 'should kill a VM if it has no checkout time' do
redis_connection_pool.with do |redis|
expect(provider).to receive(:vm_ready?).and_return(true)
expect(redis.sismember("vmpooler__running__#{pool}", vm)).to be(true)
subject._check_running_vm(vm, pool, 0, provider)
expect(redis.sismember("vmpooler__running__#{pool}", vm)).to be(true)
expect(redis.sismember("vmpooler__running__#{pool}", vm)).to be(false)
end
end
it 'should log a message when the machine is removed due to no active data' do
redis_connection_pool.with do |redis|
expect(provider).to receive(:vm_ready?).and_return(true)
expect(logger).to receive(:log).with('d',"[!] [#{pool}] '#{vm}' is listed as running, but has no checkouttime data. Removing from running")
subject._check_running_vm(vm, pool, 0, provider)
end
end