Twitter has been officially off Joyent since 10PM last night. This may come as a surprise to some after yesterday’s posts here and here regarding the two companies working together. Those of us at Joyent appreciate the opportunity we had to work with the talented folks at Twitter. It is a great service. We wish Twitter every continued success.
As I mentioned yesterday, Joyent is standing ready with excess free infrastructure to support Twitter through this transition in the event that they need it.
20 Comments
Wow, that came out of nowhere.
So who broke up with whom? Or was it mutual?
Sounds like somebody got dog housed last night…
I can’t say I blame them for leaving. You guys couldn’t even keep my blog up consistently.
How are they both off Joyent and using you for the Super Bowl? Or are you just picking up the slack?
I’m also curious about the details.
Well, this will be interesting for both parties.
fwiw I’m on shared and my blog’s been fine fine
seems to me that twitter’s problems have been due to their own frequent rejiggerings, and to their having to figure out scalable architecture while their app is scaling to hundreds of thousands of users. I’m an avid twitter user, but the app itself is still essentially a crash-and-burn prototype and it will continue to be an obvious ‘work in progress’ no matter who hosts it.
How long before Joyents lastest paramour Facebook files for divorce – I think the Joyent geeezers should stop crowing about how wonderful they feel their stuff is, there is sometimes a fine line between confidence and self-delusion or bragging ,you’ve had some recent issues (Bingo,Strongspace etc) and they should not be glossed over, perhaps you should have delayed
the resumption of self congratulating posts in the vein of “Billions Served”, “Joyent Powers” straight after the Bingo disk what happened post s- a little humility and reflection,and a couple of mea culpas ,and here’s our return to engineering discipline posts would have probably served you better.
Nice set of mixed signals in the blog postings. Just as reliable as Twitter’s uptime record.
I don’t get why they don’t just rewrite Twitter, it’s not like there is a significant amount of functionality to implement.
Short of a rewrite, their new XMPP endpoint may be their best bet. Send API traffic there, and build a non-blocking in-memory web front on top of that (with live updating Comet pages!).
Jon, it’s obvious to even the casual reader that the accelerators that served Twitter and continue to serve many Facebook sites are a completely different architecture than Bingo and Strongspace.
The problem that existed with those have absolutely nothing to do with the instability at Twitter. It’s obvious, considering the huge number of problems they’ve had today, that they have issues that have nothing to do with infrastructure.
I imagine the team at Joyent has learned a huge amount about how to design Rails apps which can scale.
Can some of this be shared? Those of us building systems are interested. Those of us intending to deploy WITH JOYENT are VERY interested.
thanks!
Fitzage, are we talking about the same post, mines the one with no mention of architectures, twitter, accelerators or instability at twitter.Mine was the one pointing out that in light of recent issues , it might be prudent to lay off the “talking up” and you know quitely excel at keeping everyone happy for a while.
Yeah, Jon, you’re the one who’s way off topic just so you can get in your bitching. You’re the one I’m talking about.
Jon: Their explanation and compensation plans regarding the Strongspace and Bingodisk downtime are honest, frank and transparent.
“Gloss over” is the worst description you could ever come by.
jgwong, Sorry , yes ,I agree with you , on re-reading it “Gloss Over” was a poor choice of words,
I’ll try again , but first some background, several years ago I worked for a large company in Australia , and its outages, often made the front page of one of Australia’s National newspapers weekly Technology editions, we were under tremendous pressure to stabilize the product, and we had to be very careful with what we communicated, as with outages or customer defections if you have a couple in a row, things can rapidly turn hostile in your business, and you end up on the back foot defending ALL your product offerings against charges that your are “Generally Unstable as a Service provider” . So I am mindful or what Joyent is going through
What I was trying to say is that Joyeur could easily have a bunch of bad new stories snowball, even if as fitzage has mentioned the offerings are unrelated to previous bad news stories, so perhaps it would be prudent to communicate messages of stability, reassure people, as sometimes posting the “Super Positive and talking up” stories at a time of uncertainty can set you up for a fall.
I apologize if I phrased it in an intemperate or overly glib manner.
I want Dean Allen back.
The details : http://reddit.com/r/programming/info/679vk/comments/c031xn8
With over 500,000 users, we’re also running one of the larger Rails applications around (CatalogChoice.org), and also grapple (well so far, thank goodness!) with scaling issues. We wrote a bit about it here:
http://summit.makalumedia.com/2008/01/28/catalog-choice-registers-half-a-million-users/
(As a side note, we originally tried to contact Joyent to discuss potential interest in hosting, and they didn’t even bother replying to us.
Is it me or is Joyent starting to sound more and more like Microsoft (if you haven’t read it yet go read http://www.simple-talk.com/community/blogs/philfactor/archive/2008/01/27/43174.aspx , it’s brilliant).
Enough of the “it’s amazing how good and scalable our system is” and “we’re so good it hurts” crap…
All I was asking from you was to keep my tiny little blog up but you couldn’t (load of 60 anyone?) so I’m out of here…
Sanjay: we intend to share more ‘best practices’ here on the blog and I am working with Jason (CTO) on that now. Stay tuned.
P.S. I am also locking down comments on several of the posts as we seem to be on a spammers hit list right now.
2 Trackbacks/Pingbacks
[...] [33] Twitter stopped using Joyent’s cloud at 10PM, Jan 30, 2008. (http://www.joyent.com/joyeurblog/2008/01/31/twitter-and-joyent-update/) [34] The hasty divorce for Twitter and Joyent. [...]
[...] [33] Twitter stopped using Joyent’s cloud at 10PM, Jan 30, 2008. (http://www.joyent.com/joyeurblog/2008/01/31/twitter-and-joyent-update/) [34] The hasty divorce for Twitter and Joyent. [...]