htrace-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Billie Rinaldi <bil...@apache.org>
Subject Re: incubator report / podling status
Date Tue, 07 Mar 2017 20:20:15 GMT
On Tue, Mar 7, 2017 at 9:11 AM, Colin McCabe <cmccabe314@fastmail.com>
wrote:

> On Mon, Mar 6, 2017, at 10:56, Billie Rinaldi wrote:
> > On Fri, Mar 3, 2017 at 8:21 AM, Stack <stack@duboce.net> wrote:
> >
> > > On Fri, Mar 3, 2017 at 4:13 PM, Billie Rinaldi <billie@apache.org>
> wrote:
> > >
> > > > I posted the report to https://wiki.apache.org/incubator/March2017
> and
> > > > signed off. Other mentors, please sign off as well.
> > > >
> > > >
> > > Done
> > >
> > >
> > >
> > > > There has been some discussion in the Accumulo community about
> > > contributing
> > > > its span receiver back to HTrace. We know it works well, and that
> would
> > > > make configuring and versioning it easier. I haven't had time to
> work on
> > > > it, though, which is why I haven't brought it up. Would HTrace
> consider
> > > > accepting an Accumulo span receiver, if we could find someone to
> work on
> > > > it?
> > > >
> > > >
> > > Yes. Makes sense. What is special about this receiver?
> > >
> >
> > Nearly all Apache Accumulo instances are configured to send traces to
> > themselves using this mechanism. There's a tracer process that spans are
> > sent to via Thrift, and the tracer inserts the spans into Accumulo. The
> > tracer location(s) are discovered via ZooKeeper. So, pretty normal?
> > https://github.com/apache/accumulo/tree/master/server/tracer
> /src/main/java/org/apache/accumulo/tracer
>
> Yeah, we should add that SpanReceiver to HTrace.
>
> >
> > When testing out HDFS and Accumulo tracing together, I found I had to add
> > a
> > configuration property that would optionally allow the receiver to drop
> > spans of length 0ms, as HDFS was generating a huge number of these. (I
> > realize dropping spans is not recommended, but I was still able to do
> > some
> > interpretation of parentless spans when the 0ms spans were missing.)
>
> This should be fixed in the latest HDFS.  Unfortunately, the Hadoop 2.8
> release got eaten by a grue, but if you try Hadoop 3.0 alpha, it should
> be there.  Incidentally, updating Accumulo to HTrace 4.x would be one of
> the best things we could do to help HTrace...
>

We've had a discussion about that. I think it will happen when we bump our
default Hadoop version to one that depends on HTrace 4. It will mean
dropping support for Accumulo+HDFS tracing for versions of Hadoop using
HTrace 3, but we decided that would be preferable to having a shim layer.
Another reason it would be nice to have the Accumulo receiver in HTrace.


>
> cheers,
> Colin
>
> >
> > Billie
> >
> >
> > >
> > > St.Ack
> > >
> > >
> > > > On Thu, Mar 2, 2017 at 4:54 PM, Colin McCabe <cmccabe@apache.org>
> wrote:
> > > >
> > > > > On Thu, Mar 2, 2017, at 13:01, Stack wrote:
> > > > > > On Thu, Mar 2, 2017 at 8:19 PM, Billie Rinaldi <
> billie@apache.org>
> > > > > wrote:
> > > > > >
> > > > > > > HTrace's quarterly report is slightly overdue (due yesterday).
> I've
> > > > > pasted
> > > > > > > a bare bones draft below based on the last report. Does
anyone
> have
> > > > > > > something to add?
> > > > > > >
> > > > > > > Thank you Billie. +1 on the report. It conveys that little
has
> > > > happened
> > > > > > since our last report.
> > > > > >
> > > > > >
> > > > > > > We should have a discussion about HTrace's status. We were
> talking
> > > > > about
> > > > > > > the possibility of graduation over a year ago. Do you think
we
> need
> > > > to
> > > > > grow
> > > > > > > the community further before proposing graduation?
> > > > > > >
> > > > > > >
> > > > > > We should. There has been little progress of late. Lets talk
> about
> > > > this.
> > > > > > Most of us seem to be off occupied on other stuff with little
> time
> > > over
> > > > > > for
> > > > > > moving HTrace forward.
> > > > > >
> > > > > > St.Ack
> > > > >
> > > > > Thanks for speaking up, Billie and Stack.  Agree it's been a little
> > > > > quiet lately.  I've been busy on Apache Kafka recently.  There was
> some
> > > > > stuff I wanted to push on the HTrace side but I didn't get a
> chance to
> > > > > post it yet.
> > > > >
> > > > > Colin
> > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > -----
> > > > > > > HTrace is a tracing framework intended for use with distributed
> > > > systems
> > > > > > > written in java.
> > > > > > >
> > > > > > > HTrace has been incubating since 2014-11-11.
> > > > > > >
> > > > > > > Three most important issues to address in the move towards
> > > > graduation:
> > > > > > >
> > > > > > >   1. Continue to grow the Apache HTrace community
> > > > > > >   2. Continue to explore projects integrating Apache HTrace
> > > > > > >   3. Continue to develop and release stable Apache HTrace
> > > incubating
> > > > > > > artifacts
> > > > > > >
> > > > > > > Any issues that the Incubator PMC (IPMC) or ASF Board
> wish/need to
> > > be
> > > > > > > aware of?
> > > > > > >
> > > > > > >   There are no issues requiring board attention at this
time.
> > > > > > >
> > > > > > > How has the community developed since the last report?
> > > > > > >
> > > > > > >   Mailing lists have been quiet.
> > > > > > >
> > > > > > > How has the project developed since the last report?
> > > > > > >
> > > > > > >   Development has been quiet.
> > > > > > >
> > > > > > > How would you assess the podling's maturity?
> > > > > > > Please feel free to add your own commentary.
> > > > > > >
> > > > > > >   [ ] Initial setup
> > > > > > >   [ ] Working towards first release
> > > > > > >   [x] Community building
> > > > > > >   [ ] Nearing graduation
> > > > > > >   [ ] Other:
> > > > > > >
> > > > > > > Date of last release:
> > > > > > >
> > > > > > >   2016-10-06
> > > > > > >
> > > > > > > When were the last committers or PPMC members elected?
> > > > > > >
> > > > > > >   2016-10-03
> > > > > > >
> > > > >
> > > >
> > >
>
> --
>   Colin McCabe
>   cmccabe314@fastmail.com
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message