search

intel-cloud / cosbench

a benchmark tool for cloud object storage service
Other
573 stars 242 forks source link

stage3 always running and seemes can not stop #369

Open kun95 opened 6 years ago

kun95 commented 6 years ago

Hi, I found that cosbench seems always running and can not stop with some specific settings.

following is my workload config file:

<?xml version="1.0" encoding="UTF-8" ?>

what we want to do is to test oss by cosbench in 24 hours.

but , after stage-3 running 24 hours. it seems can not stop . and can not step into next stage.

by the way . 1.we have developed an adaptor to support aliyun oss , and the test based on this adaptor cosbench-oss_0.4.2.0.jar 2.It works well when we set runtime 、workers、sizes to a smaller value , for exemaple runtime = 200 workers=8 and sizes=10MB .

now, i am so confused about this . I dont know how to fix it.

hope you can give us some advices...

log files and adaptor

system.log workload.log workload-config.xml driver-boot.log cosbench-oss_0.4.2.0.jar

result:

MostafaNabi commented 6 years ago

Looking at your log files there seem to be some errors caused by connections timing out. Open your log file and search for 'error'. I'm not expert (just a COSBench user) but it may be caused by the large number of workers you are using. I saw a similar issue when I used 512 workers. The server may not be able to cope with having 300 workers so when it tries to serve them all it causes timeout errors. This is why when you lower your number of workers it works fine.

You can use more workers for smaller objects because the server does less work per object so it can serve more. Just noting though, that this is simply my theory, I have limited experience in this.