[kata-dev] 回复: 回复: 回复: rust kata-agent status update

杨波 yb203166 at antfin.com
Fri Sep 20 01:54:40 UTC 2019


I think the extra money is spent in go  runtime, probably.


------------------------------------------------------------------
发件人:Dr. David Alan Gilbert <dgilbert at redhat.com>
发送时间:2019年9月19日(星期四) 16:28
收件人:杨波(兵卫) <yb203166 at antfin.com>
抄 送:"Boeuf, Sebastien" <sebastien.boeuf at intel.com>; Samuel Ortiz <sameo at linux.intel.com>; kata-dev <kata-dev at lists.katacontainers.io>
主 题:Re: [kata-dev] 回复: 回复: rust kata-agent status update

* 杨波 via kata-dev (kata-dev at lists.katacontainers.io) wrote:
> go-agent: about 11M 
> rust-agent: about 1.1M

Do you understand why? That's a big difference - I wonder where the
extra memory goes?

Dave

> It is also in README.md at here: https://github.com/alipay/kata-rust-agent/blob/master/README.md
> 
> 
> 
> 
> 
> ------------------------------------------------------------------
> 发件人:Boeuf, Sebastien <sebastien.boeuf at intel.com>
> 发送时间:2019年9月19日(星期四) 05:28
> 收件人:杨波(兵卫) <yb203166 at antfin.com>; Samuel Ortiz <sameo at linux.intel.com>
> 抄 送:kata-dev <kata-dev at lists.katacontainers.io>
> 主 题:RE: [kata-dev] 回复: rust kata-agent status update
> 
> Hi,
> 
> I love the footprint drop due to using Rust :)
> Instead of giving only the numbers from the Rust agent, could you show the difference between the Golang agent and Rust agent?
> 
> Thanks,
> Sebastien
> 
> From: 杨波 via kata-dev [kata-dev at lists.katacontainers.io]
> Sent: Sunday, September 15, 2019 6:52 PM
> To: Samuel Ortiz
> Cc: kata-dev
> Subject: [kata-dev] 回复: rust kata-agent status update
> 
> footprint:
> RssAnon: 1700 kB
> RssFile: 1632 kB
> 
> We choose grpc-rs because it is verified in product environment, grpcio is not.
> 
> 
> 
> ------------------------------------------------------------------
> 发件人:Samuel Ortiz <sameo at linux.intel.com>
> 发送时间:2019年9月12日(星期四) 20:44
> 收件人:杨波(兵卫) <yb203166 at antfin.com>
> 抄 送:kata-dev <kata-dev at lists.katacontainers.io>; 王旭(循环) <gnawux at alibaba-inc.com>
> 主 题:Re: [kata-dev] rust kata-agent status update
> 
>  Hi,
> 
>  A couple of questions:
> 
>  1) Do we know what the grps-rs memory footprint is?
>  2) What was the rationale behind picking grps-rs over e.g. grpcio?
> 
>  Cheers,
>  Samuel.
> 
>  On Thu, Sep 12, 2019 at 04:53:29PM +0800, 杨波 via kata-dev wrote:
>  > Hi, folks:
>  > Let me update some progress of the rust agent.
>  > First of all, thanks for James' kindly help on the project.
>  > The functions of rust agent are almost finished. we have fixed several bugs to stablize it. Ci and osbuilder are still missing. We have to run end-to-end tests manually right now.
>  > 
>  > - Refactor rust agent API. Remove overused matches, clean up code
>  > - Refactor communication between parent/child/grandchild, use raw bytes instead of serde_json to simplify communication
>  > - Refactor Process struct, remove duplicate/unused fields
>  > - Add Makefile to simplify building --- by jodh-intel
>  > - Fix up unit test for netlink
>  > - Fix up test run against musl and release
>  > - Implement debug console support
>  > - Implement command line parse partly
>  > - Implement Spec valaidator
>  > - Implement Hook infrastructure
>  > 
>  > We are targeting to land the rust agent to Kata org by October, and you may check out the current WIP repo here: https://github.com/alipay/kata-rust-agent
>  > 
>  > Cheers!
>  > 
> 
>  > _______________________________________________
>  > kata-dev mailing list
>  > kata-dev at lists.katacontainers.io
>  > http://lists.katacontainers.io/cgi-bin/mailman/listinfo/kata-dev
> 
> 

> _______________________________________________
> kata-dev mailing list
> kata-dev at lists.katacontainers.io
> http://lists.katacontainers.io/cgi-bin/mailman/listinfo/kata-dev

--
Dr. David Alan Gilbert / dgilbert at redhat.com / Manchester, UK
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.katacontainers.io/pipermail/kata-dev/attachments/20190920/5bd1fdf9/attachment.html>


More information about the kata-dev mailing list