Untangle basho_stats_writer and update travis

basho / basho_bench

A load-generation and testing tool for basically whatever you can write a returning Erlang function for.

Apache License 2.0

311 stars 197 forks source link

Untangle basho_stats_writer and update travis #192

Closed linearregression closed 8 years ago

linearregression commented 8 years ago

untangle basho_stats_writer. csv and riemann writer no longer in one file.
update travis.yml (at least one of pulled deps now require min erlang R16). Use travis container to build.

slfritchie commented 8 years ago

Hi, sorry about the delay in reviewing this PR!

The R16 minimum runtime requirement of the quickrand module is flawed. As far as I can tell, the random_wh06_int:uniform(5316911983139663491615228241121378304) function works just fine when using R14B04, if I cut out the cruft that keeps quickrand from compiling at all.

^C250 [sbb4]:quickrand: env PATH=/usr/local/erlang/R14B04.64bit/bin:$PATH erl -pz ebin
Erlang R14B04 (erts-5.8.5) [source] [64-bit] [smp:8:8] [rq:8] [async-threads:0] [hipe] [kernel-poll:false]

Eshell V5.8.5  (abort with ^G)
1> random_wh06_int:uniform(5316911983139663491615228241121378304).
21209109798739021856976652074467755

I'm not a fan of forcing an Erlang runtime dependency change for basho_bench. Do you have an alternate recommendation?

linearregression commented 8 years ago

The last travis CI build was ran 2yrs ago: https://travis-ci.org/basho/basho_bench/branches It does look like this repo is not active.

The majority of the PR is to untangle to stats backend (.e.g you merged graphite and one other as metrics reporting backend).

So if you can make travisci run good again with R14 to R18. I can just merge with your branch or .master, then there is no need to remove R14 support.

slfritchie commented 8 years ago

{exasperation} ... the katja library uses a function name query(), which is a reserved word in Erlang R15.

slfritchie commented 8 years ago

See also, this commit after I merged this branch: https://github.com/basho/basho_bench/commit/f941cfd53184795561b4b299efbf89cfb6d192cc

Many thanks for the PR!

linearregression commented 8 years ago

Have you considered demoting the module using kaja as contrb("use as at your own risk")? The statsbackend is now untangled.

It does not look like katja is active in either branches. You just want to post stats to Riemann server; does not look like you want any special functionality from the library. It's protocolbuff deps is less than ideal.

https://github.com/nifoc/katja/tree/development

On Tue, Dec 8, 2015 at 6:25 PM, Scott Lystig Fritchie < notifications@github.com> wrote:

See also, this commit after I merged this branch: f941cfd https://github.com/basho/basho_bench/commit/f941cfd53184795561b4b299efbf89cfb6d192cc

Many thanks for the PR!

— Reply to this email directly or view it on GitHub https://github.com/basho/basho_bench/pull/192#issuecomment-163086512.

slfritchie commented 8 years ago

Riemann is not my circus nor katja my monkey, sorry If you have a PR in mind, I'll be happy to review it, but someone who's familiar with Riemann and the Riemann client library ecosystem should take a shot at this. If it were me, I'd rip it out, so don't let me do it. ^_^