 | |
| saw that? pretty incredible. |
| |
| |
| it's almost the only point with Dynamo systems, otherwise why to use such a system if the result is data loss? |
| |
| |
| shopping cart? You want a Set performing union. Time series? You want a list performing smart head/tail merging, and so forth. |
| |
| |
| so instead of dealign client-side with multiple returned vales, you pick the data structure depending on what you want. |
| |
| |
| strategy for the different types. Like sets will do unions, and so forth. Maybe strings will just use clocks instead. |
| |
| |
| an idea I explored time ago: take a dynamo system and put Redis-alike data structures on top of it. Then set a default merging... |
| |
| |
| IMHO the most shocking thing in the series is to see how with wrong defaults Dynamo systems like Riak become fragile. |
| |
| |
| good idea, worth to try. Still this should not be a problem at a first glance, clock jumps are much worse in general. |
| |
| |
| allow to take the time unaffected of the clock, but guaranteed to monotonically increase. This could be useful. |
| |
| |
| what you can cause is to change the node timeout from the point of view of some Sentinel. Btw in Linux there are sys calls that |
| |
| |
| if you skew the clock continuously but very little every second, you can do that, but this will harm very little AFAIK (cont) |
| |
| |
| ciao Romero, basta la documentazione su redis.io. Potresti anche leggere qui: |
| |
| |
| thank you, not due, but appreciated. The only factual incorrect thing AFAIK is the TILT stuff, assuming I understood what you meant. |
| |
| |
| I'm sure about that but the community not caring about high quality work is not a good sign. |
| |
| |
| It is incredible to me that none of the blog posts hit the HN front page yesterday. Hope my post will help popularize his work. |
| |
 | |
| “If a picture is worth 1000 words, a prototype is worth 1000 meetings" |
| |
| |
| my reply to article about Sentinel -> |
| |
| |
| I'm not defensive, I'm glad to him honestly. |
| |
| |
| Look at how attacks many data stores, including Redis => That was a great work. I'm writing my reply right now |
| |
 | |
| Seriously read this. Learn about data loss with Postgres, ACID & Two-phase commit. RT : Call me maybe: Postgres |
| |
