Re: use Jaeger tracing instead of zipkin in librados

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



hey Omri (and cc dev list),

On Mon, Jun 13, 2022 at 6:51 AM Omri Zeneva <ozeneva@xxxxxxxxxx> wrote:
>
> Hi everyone,
>
> I started my work on creating a single jaeger trace which starts on the RGW and goes down to the rados. I'm currently investigating how can I pass the info of the trace down to the rados,  in order to create a child span of this trace.
> There is already a blkin trace info object in some functions in librados, for example:
> https://github.com/ceph/ceph/blob/main/src/librados/IoCtxImpl.h#L159-L163
> which is used for the Zipkin tracing.
>
> I'm wondering if this Zipkin tracing is actually in use, and if we can use this param for the jaeger trace info (actually opentelemetry's trace info, which also supports Zipkin as its tracing system)
> do you think it is feasible? or it will be better to leave it there?

i would love to start removing the zipkin tracing stuff. it should be
safe to remove it from the client side (librados and Objecter), and
just replace their 'blkin_trace_info*' arguments with an opentracing
type

we'll just need to be careful about backward compatibility, because
blkin_trace_info is encoded in the messenger protocol - see
Message::encode/decode_trace(). on the server side, we'll need to
continue supporting the decode of blkin_trace_info sent by older
clients. but past that, the messenger and OSD can probably ignore
blkin traces entirely and only process the opentracing ones

_______________________________________________
Dev mailing list -- dev@xxxxxxx
To unsubscribe send an email to dev-leave@xxxxxxx



[Index of Archives]     [CEPH Users]     [Ceph Devel]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux