Re: Performance tuning of ServiceComb Saga Pack

2018-08-20 Thread Zheng Feng
2018年8月20日(星期一) 中午11:15 > 收件人: "dev"; > > 主题: Re: Performance tuning of ServiceComb Saga Pack > > > > We could use the async writer to store the events to the database when > using the redis. Redis cluster has the persistent storage at the same time. > > > > Will

?????? Performance tuning of ServiceComb Saga Pack

2018-08-19 Thread bismy
the really important data is persistent. -- -- ??: "willem.jiang"; : 2018??8??20??(??) 11:15 ??: "dev"; ????: Re: Performance tuning of ServiceComb Saga Pack We could use the async writer to store the events to the

Re: Performance tuning of ServiceComb Saga Pack

2018-08-19 Thread Willem Jiang
clusters and sync memory data between > clusters and assume that clusters is high available? > > > -- 原始邮件 -- > 发件人: "willem.jiang"; > 发送时间: 2018年8月16日(星期四) 中午11:00 > 收件人: "dev"; > > 主题: Re: Performance tuning of ServiceComb

Re: Performance tuning of ServiceComb Saga Pack

2018-08-15 Thread Willem Jiang
async way.can > we guarantee update on both > db and redis are success at same time when the transation is aborted? > 发件人: Willem Jiang > 发送时间: 8月16日星期四 09:40 > 主题: Re: Performance tuning of ServiceComb Saga Pack > 收件人: dev@servicecomb.apache.org > > > Hi Amos Alpha send respo

Re: Performance tuning of ServiceComb Saga Pack

2018-08-15 Thread fu chengeng
Hi Willem why not just store 'finished transacation' data to db in a async way.can we guarantee update on both db and redis are success at same time when the transation is aborted? 发件人: Willem Jiang 发送时间: 8月16日星期四 09:40 主题: Re: Performance tuning of ServiceComb Saga Pack 收件人: dev

Re: Performance tuning of ServiceComb Saga Pack

2018-08-15 Thread Zheng Feng
Thanksk Willem, it looks like the redis cluster is the better choice. 2018-08-16 9:57 GMT+08:00 Lance Ju : > Here is an abstracted design of how redis works: > https://raw.githubusercontent.com/crystaldust/saga- > performance-optimization/master/saga-redis.jpg > > Willem Jiang 于2018年8月16日周四

Re: Performance tuning of ServiceComb Saga Pack

2018-08-15 Thread Lance Ju
Here is an abstracted design of how redis works: https://raw.githubusercontent.com/crystaldust/saga-performance-optimization/master/saga-redis.jpg Willem Jiang 于2018年8月16日周四 上午9:40写道: > Hi Amos > > Alpha send response to the Omega once the message is updated into redis, > then we just store the

Re: Performance tuning of ServiceComb Saga Pack

2018-08-15 Thread Willem Jiang
Hi Amos Alpha send response to the Omega once the message is updated into redis, then we just store the transaction events into the database in async way (we don't change the states here). Current Redis cluster provides the persistent storage, it could reduce lot of effort of us. Now we just use

Re: Performance tuning of ServiceComb Saga Pack

2018-08-15 Thread Zheng Feng
id<https://aka.ms/ghei36> > > > From: Zheng Feng > Sent: Wednesday, August 15, 2018 10:48:34 PM > To: dev@servicecomb.apache.org > Subject: Re: Performance tuning of ServiceComb Saga Pack > > Hi Willem, > > It makes sense to use the redis to store the pending tran

Re: Performance tuning of ServiceComb Saga Pack

2018-08-15 Thread fu chengeng
.ms/ghei36> From: Zheng Feng Sent: Wednesday, August 15, 2018 10:48:34 PM To: dev@servicecomb.apache.org Subject: Re: Performance tuning of ServiceComb Saga Pack Hi Willem, It makes sense to use the redis to store the pending transactions (I assume that yo

Re: Performance tuning of ServiceComb Saga Pack

2018-08-15 Thread Zheng Feng
Hi Willem, It makes sense to use the redis to store the pending transactions (I assume that you mean these are the "HOT" ones). But we could be very careful to "write" the transaction status, and it should be stored in the database at last. So I think we must make sure the transaction status in

Performance tuning of ServiceComb Saga Pack

2018-08-14 Thread Willem Jiang
Hi, With the help of JuZheng[1][2], we managed to deploy the saga-spring-demo into K8s and start the Jmeter tests for it. By running the test for a while, the DB CPU usage is very high and the response time is up 2~3 seconds per call. It looks like all the event are stored into the database in