amazon web services - AWS EC2 (batch) throttling connection? -
so running script on aws batch fetches 100,000+ domains.
- the script runs in docker container.
- the domains randomized , in redis queue.
- the script pulls 20,000 domains queue, processes them , writes results redis again.
when run script on ec2 instance launched get.
ami: custom ebs optimized: true time elapsed : 2007.7330884933472 domains : 53517 processed domains: 240000 when run on ec2 instance spun batch get:
ami: amazon linux ami 2017.03.e x86_64 ecs hvm ebs optimized: false time elapsed : 2313.34757232666 new domains : 51243 processed domains: 400000 are instance's aws batch spins throttling connection? because using docker can't think of why results different other bandwidth issues. docker image stored in ecs , pulled down , script run.
i have run test on millions of randomized domains , results same statistically speaking not sampling of domains either. domain rate 2x on instance spin vs batch.
update 1: differences, ebs true on 1 , false on other. can't seem change aws batch.
update 2: tested ebs false machine lower specs , still can't account lower network performance. maybe ami?
update 3: tested ami , problem ami-c6f81abe used on batch. not sure why yet.
update 4: turns out ulimit parameter on jobdefinition causing problem.
Comments
Post a Comment