Is there high avalibility mechnisms for worker instances?


#1

Does Scidb worker instances replicate data to somewhere else in case of worker nodes failure, and does one worker instance failure cause the whole custer stop?


#2

Hello,

The SciDB Community Edition system does not support this. The Paradigm4 SciDB enterprise edition does have a system plugin that allows you to set the system up for k-replication. If used, the system will stay live and data will remain readable after up to k-1 instance failures. To find out more about it, contact Paradigm4 via paradigm4.com/product/try-it-out/

  • Alex Poliakov

#3

Hi, do you mean the “redundancy (optional)” parameter, and do you mean we can’t use this parameter in the community edition of SicDB?


#4

I read the edition comparation page at paradigm4.com/licenses/, it says k- replication feature is avaliable at both editions.


#5

So how can we recover from a worker instance or worker server failure, and the cordinator instance or server?


#6

The current version of the page that compares CE vs. EE (paradigm4.com/product/try-it-out/