Replies: 1 comment 2 replies
-
Hi @rez0n It’s good that there is no requirement for write speed in favor of data safety.
And, if you only have two replicas:
For more confidence, it is desirable:
but this is possible if you have at least 3 replica servers in the Cluster since writes will not be possible if less than 2 replica servers are available. I recommend using synchronous replication in a production environment if possible. Details: |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi there,
I choosing between synchronous or asynchronous mode of 3 node cluster (+2 external etcd) for a low count of writes database (it can be represented as typical e-commerce website)
I need maximum stability (writes & reads) and performance (reads)
I've tested asynchronous mode by making few inserts and disconnecting network of primary node but not faced with any data loss. I still unsure, how much risk of data loss in practical point of view, it becomes critical on high level of writes or it's random?
In addition, I not expect frequent failovering by real outage, most of them should to be due software issues or reboot (I have issue in my testing lab with random reboots by watchdog, most probably because I reduced patroni_ttl without proper understanding what is it for)
Thanks in advance.
Beta Was this translation helpful? Give feedback.
All reactions