skyde

46 ms·

1. 511.▲
   by skyde 6y ago

   Pulsar itself is solid but it’s true number of good clients is limited. Same is true for kafka the only reliable client is the java one. All other are based on librdkafka which still have many major bugs causing it to be stuck and not produ
2. 512.▲
   by skyde 6y ago

   wow, do you remember which bank and as the backend for what kind of system ? if it’s just for some kind of cache or system that is ok to randomly corrupt and drop data I would sleep better :-)
3. 513.▲
   by skyde 6y ago

   Anyone know how large scale chat system ( facebook messenger, ...) are implemented? My guess is message are in a data store like hbase and a very simple notification system let user that are online know to fetch for new entry
4. 514.▲
   by skyde 6y ago

   Are you saying GCP pubsub have bad reliability? I never used it but would be surprised if it was down for several hours each weak or would just drop your data. I would love to know your experience
5. 515.▲
   by skyde 6y ago

   MongoDB is still pretty much broken. I really hope your bank is not using MongoDB.
6. 516.▲
   by skyde 6y ago

   exactly once messaging is a lie. it just mean you have duplicates message but let Kafka try to detect duplicate for you, which is almost never what you want. it’s much simpler to make your message processing indempotent.
7. 517.▲
   by skyde 6y ago

   my experience is the opposite unless you really need the whiz-bang feature you should never use Kafka it’s the least reliable and hardest to run it require active babysitting by skilled team of admin and ops. if you don’t need low end-to-e
8. 518.▲
   by skyde 6y ago

   Kafka is not more mature just more hyped. I just wish Aphyr jepsen test would also cover more scenario like - what happen to your data if x+1 server permanently fail in the cluster with a replication factor of X. - what happen if a single p
9. 519.▲
   by skyde 6y ago

   The main aspect for me after 5 years of running several Kafka cluster to run production critical system at Microsoft is that it’s very fragile because of the bad design. distributed system are hard to design correctly. When using HDFS you
10. 520.▲
    by skyde 6y ago

    While I agree I see a big problem that people often pick Kafka when they really needed Rabbitmq because of the hype around Kafka and not understanding the difference between a transactional message broker ( Rabbitmq and ibm mq ...) as used
11. 521.▲
    by skyde 6y ago

    Pulsar can also support infinite data retention using data tiering and by spreading data on all nodes. It’s much better than Kafka when you don’t know when the consumer will come back looking for what changed
12. 522.▲
    by skyde 6y ago

    If you need. Ordered log abstraction (Apache pulsar, Facebook log device ...) if you need a real transactional and highly available message broker any (my/jms) server like Rabbitmq IBM MQ .... Kafka in my experience is always the worse
13. 523.▲
    by skyde 6y ago

    What every engineer should know about Kafka is that it should not be used for anything critical like you would use Cassandra or Hbase. But if you are ok with partitions not being available for many hours or losing all written data because t
14. 524.▲
    by skyde 6y ago

    agree with you (computerGuru) at large scale any rare or corner case happen multiple time per day :-) Redis was never designed for strong Synchronous replication of transaction or Strong leader election. So each time a master crash you have
15. 525.▲
    by skyde 6y ago

    Only work starting windows 10 ?
16. 526.▲
    by skyde 6y ago

    will it be possible to use the « sendfile » system call to do zero copy file transfers on a quick connection?
17. 527.▲
    by skyde 6y ago

    is this using BBR congestion algorithm ?
18. 528.▲
    by skyde 6y ago

    will this be integrated in some rpc framework like gRPC we all can use ?
19. 529.▲
    by skyde 6y ago

    IPFS need a way to get paid for a block over a period of time. Does not need to be real money just token you can use later to pay for your file to be seeded. Basically I host your file and you host my file. With some algorithm to make sure
20. 530.▲
    by skyde 7y ago

    can someone explain how persistence work in swimOS. Some kind of distributed database running on each node ?
21. 531.▲
    by skyde 7y ago

    why use iso 8,000 as input and not the camera native ISO ?
22. 532.▲
    by skyde 7y ago

    would have loved to see raid-Z in this test.
23. 533.▲
    by skyde 7y ago

    Any one know why the Turbo Boyer-Moore algorithm is not more popular. Any case when Boyer Moore would be better?
24. 534.▲
    by skyde 7y ago

    This guy is a hero! I have so much respect for his work. By i’m still wondering if this bug could have been detected by automatic fuzzing testing ?
25. 535.▲
    by skyde 7y ago

    Why do you think Rust would be different from Java or C#. I believe that both Java and C# are doing a good job at sharing package using Maven and Nuget.
26. 536.▲
    by skyde 7y ago

    thats why Rust might win in the long run because it have a common build system and package manager
27. 537.▲
    by skyde 7y ago

    What about WD Gold drives ? are they also using shingled magnetic recording
28. 538.▲
    by skyde 7y ago

    spark version of Ada is such a language but it use design by contract assertion only.
29. 539.▲
    by skyde 7y ago

    is it one postSQL per app replacing many app on a single Oracle instance ? moving away from some sharing fix most of the problem DBA are hired to mitigate
30. 540.▲
    by skyde 7y ago

    you can have a kernel thread pool dedicated to one kernel feature. This way only task created by this kernel feature will run on this pool and compete together .