WebScaleSQL: A collaboration to build upon the MySQL upstream | Hacker News Reader

WebScaleSQL: A collaboration to build upon the MySQL upstream

code.facebook.com

108 pointsmikeevans28 comments

From Twitter:

https://twitter.com/TwitterOSS/status/449202984786739200 https://github.com/twitter/mysql/wiki/WebScaleSQL-and-Twitte...

Doesn't Percona's MySQL "fork" already include most of the Twitter, Google and Facebook patches?

They've included only one of the Twitter patches that I know of -- Jeremy Cole's NUMA pre-allocation and interleaving work. Percona contributes a lot of great stuff to the MySQL community, but their focus has always been supporting their paying customers.

I wonder how Google's broad adoption of MariaDB figures into this news.

Isn't the mysql test suite no longer open source/available ?I thought oracle decided to hide them ?

The test suite is still there. I have had questions about the number of new tests that are not getting published -- http://mysqlha.blogspot.com/2013/11/where-are-tests.html

I wonder if it is because of the new pending comments system that 15 comments in, no one has said, "why don't they just use PostgreSQL?" :-). Not that I am MySQL fanatic but I think it gets more hate on HN than it deserves. It has its quirks but is a good lightweight DB.

Funny enough I think all the hate for MySQL provides good documentation of what to watch out for.

MariaDB, webscale, Percona... it's really annoying all those forks, why can't they choose one solution to work on together.

WebScale is supposed to be a branch, not a fork but since you asked:

Percona built a business on 'their' version and operate on the RedHat business model. As such they need a blessed copy for their business model to function.

MariaDB is basically a bid to take control of MySQL back by one of the original MySQL developers.

Etc.

Tbh, the main problem is Oracle has been a terrible steward of MySQL which has led to this mess.

Percona is no more a fork than Webscale is, it's supposed to be a drop-in replacement for Oracle MySQL and it tracks releases pretty closely.

MariaDB was supposed to be a drop-in replacement, but they always had much grander designs than Percona or Webscale do, so they found it was no longer in their interests to track upstream so closely. So after 5.5, MariaDB was going to jump straight to version 10.0. 5.5 is still the last stable release of MariaDB, though, even though Oracle MySQL hit 5.6 over a year ago.

I think "terrible steward" is way off base. 5.5 and 5.6 have both been good releases with some pretty significant improvements for real users, and there's even more stuff coming in 5.7. My general impression is that MariaDB doesn't have the resources to keep up with what Oracle is doing technically, and I'm betting that technical merits will win out.

http://www.theregister.co.uk/2013/09/12/google_mariadb_mysql...

Also, I read somewhere that Google is contributing engineers to work on Maria. Although how many and for how long (compared to the GOOG's internal databases) remains to be seen.

Google is contributing code to WebscaleDB too, I don't know that their support of MariaDB means anything in this context. If you want MySQL+Google enhancements, WebscaleDB seems to give you that, plus the efforts of Facebook, Twitter, LinkedIn and (since Webscale is not a fork) Oracle.

I love TheReg, but I am not sure I would use them as a primary source.

I think Oracle has been a good steward. They haven't been perfect but they have done a great job at paying back technical debt and making the product better for modern hardware.

Both releases (5.5 and 5.6) since Oracle acquired MySQL have been mostly pain-free and have brought excellent scalability and performance improvements. If you remember the fiasco that Sun gave us with MySQL 5.1, I think you'd find Oracle's stewardship a welcome change.

Everyone is welcome to their opinion but, personally, I switched to Maria because I think they are better stewards of an open source project than Oracle is.

http://mysqlha.blogspot.com/2012/08/less-open-source.html etc. etc.

You are free to hold a separate and equally valid opinion. :)

I guess I am free to hold both opinions. That is my blog. As I write elsewhere on this page I think Oracle has been a good steward of MySQL, but have not been perfect. Maybe they will improve?

That would be the hope. :)

It would not be a stretch to call percona server a branch as well. It has rebased twice already (5.5 and 5.6) and tries to focus on high impact small changes.

Not to make excuses, but oracles position of supporting releases for 8 years on many platforms is not an easy one. Being able to see what works out in branches that don't have this constraint is actually great for the whole ecosystem.

(Disclaimer: former percona employee. Current oracle employee)

NB: this is a branch, rather than a fork.

But is it web scale?!

it's only web scale if you join tables directly in the PHP code!

You must be talking about RethinkDB then, that's their selling point!

It is only web scale if all the cool kids start using it. :)

It's interesting that the development is not really designed for public contributions; only contributions originating in the "webscale" companies.

There is nothing that precludes public contributions to the project. The important part is that the change be useful to 'webscale' type applications.

Prior to this announcement, the only people involved have been these companies to get it bootstrapped, so the work so far looks a bit slanted to these companies.

Taken from the FAQ:

http://webscalesql.org/faq.html

"We’ll be doing everything in the open, so everyone in the MySQL community will be able to take what they want from what we do and contribute however they like."

Is the project committed to being 100% compatible with oracle mysql?

/dev/null

Reply on news.ycombinator.com