aphyr

30 ms·

271.
by aphyr 10y ago
Clustering columns do not make Cassandra updates any safer: they only reduce the scope of conflicts. It's still last-write-wins--the approach that Cassandra's own blog recognizes as "a high potential for data loss". Tha
272.
by aphyr 10y ago
http://www.slideshare.net/NorbertoLeite/mongodb-wiredtiger-i...
273.
by aphyr 10y ago
> assume updates can be sometimes applied in a different order than they were submitted I suggest you re-read the analysis. Some databases can offer safe, generalized commutative updates; e.g. Riak. Cassandra can't: updates, in gene
274.
by aphyr 10y ago
Pretty sure Cassandra still uses local timestamps and last-write-wins, which means you can't do safe updates. Also pretty sure they didn't fix the timestamp-collision row-isolation issue I talked about, because it was on HN's
275.
by aphyr 10y ago
I'm not exactly sure what you're trying to say, but perhaps I can help provide context: In 2013, I performed an unpaid analysis, in my nights and weekends, of MongoDB. I found a bug leading to the loss of acknowledged writes with
276.
by aphyr 10y ago
It might be better to think of it as a limitation of two-generals, rather than 2PC in particular.
277.
by aphyr 10y ago
Cassandra had a minor bug with timestamps which was fixed relatively fast and now it passes Jepsen fully. Wait, what?
278.
by aphyr 10y ago
What does MongoDB bring that couldn't have 'built on top' of MySQL codebase, and used MySQL transational layer as it's underpinnings. There are two challenges here. One is that MongoDB has a different data model than M
279.
by aphyr 10y ago
Er, to be clear, they have actually fixed those issues in the v1 protocol. There may be other bugs, but I haven't been able to find them yet.
280.
by aphyr 10y ago
Thanks. :)
281.
by aphyr 10y ago
Protocol v1 should be largely transparent to clients (at least for standard kv operations). It's a change you make on the server, at the replica set level.
282.
by aphyr 10y ago
I agree. In my ideal world I publish everything immediately. There's no "standard practice" that I know of--very few people are doing this kind of work. Behind every one of these analyses is weeks of contract negotiation wher
283.
by aphyr 10y ago
While you are correct that 3.4 now passes, I feel like your interpretation here is a bit optimistic. It's not like I tested Mongo and it passed out of the box--it failed Jepsen tests, and not just with a read anomaly--it lost majority-
284.
by aphyr 10y ago
This analysis was a collaboration with MongoDB over the past three months. MongoDB was able to fix these issues in 3.4.0 because I found them in rc3 and rc4.
285.
MongoDB 3.4.0-rc3(jepsen.io)
226 points by aphyr 10y ago|162 comments
286.
by aphyr 10y ago
(I've since skimmed Charron-Bost & Cori, and it shows that linearizability is not composable when there does not exist a total order of invocation and response events. This might be of use in relativistic scenarios with... accelera
287.
by aphyr 10y ago
Most people require that linearizable execution must happen as-if it matches a global clock. Simply requiring that the effects of a transaction happen within the start/end time of the partition executing the transaction does not guara
288.
by aphyr 13y ago
On the contrary, I believe strongly in the importance of allies.
289.
by aphyr 13y ago
I want to help other groups in this industry; however, there is no speaking to them. I think I partially know how it feels to be gay, because I didn't choose my situation yet there is an extremely vocal group that wants me to believ
290.
by aphyr 13y ago
Well this one's proof by construction, so I don't feel the need to go particularly deep into the math. If you prefer a more rigorous approach take a look at the TLA proofs earlier in the Jepsen series; think I linked one on Redis+
291.
by aphyr 13y ago
I assert that this design actually allows for arbitrarily long windows of data loss, but I haven't verified that the implementation matches my understanding of antirez's WAIT/failover algorithm yet. Pretty sure though.
292.
by aphyr 13y ago
Today I learned that I'm straight, Nancy Lynch is a man, and David Nolen is white. Diverse engineers are all around us. They just go unnoticed--in part, because we interact with usernames and papers instead of face to face.
293.
by aphyr 13y ago
It's not necessary for all development, but you'll want a java compiler for later parts of the guide.
294.
by aphyr 13y ago
Three reasons. First, I want to justify and contextualize the work in the hopes of promoting more like it. Second, stereotype threat is ameliorated by explicit acknowledgement of the effect. Third, I want to challenge implicit assumptions a
295.
by aphyr 13y ago
Thank you, and yes, there will be more.
296.
by aphyr 13y ago
The linked leiningen page includes detailed installation instructions for all platforms, including a windows installer.
297.
by aphyr 13y ago
WriteNumbers are a logical clock provided by (last I checked) a single coordinating HBase node, which is a totally valid and commonly used way to order operations. For the purposes of this article, though, I'm trying to focus on LWW da
298.
by aphyr 13y ago
It's a complex problem. The best overview I know of is here: http://www.ucolick.org/~sla/leapsecs/onlinebib.html
299.
by aphyr 13y ago
Once an event is visible to all users, There's the rub: any GC/compaction (and when you get down to it, any read ) is actually a distributed consensus problem. If you're interested in this problem, you might take a look at
300.
by aphyr 13y ago
Yes.

More ›