Replies: 4 comments 2 replies
-
i got it working with 2 nodes in the past using the command line version, just need to make 2 node etcd cluster, however this means that you wont be able to achieve any failover with just 2 etcd nodes because it will go into read only mode if only 1 is online . etcd requires 3 for failover , or 2 running to write/read. |
Beta Was this translation helpful? Give feedback.
-
For a high-availability PostgreSQL cluster, at least 2 servers are required (Primary + Replica). However, for the DCS cluster (etcd or Consul) used by the PostgreSQL cluster, at least 3 servers are necessary to ensure quorum (see the implementation of the RAFT algorithm). Alternatively, you can start with a single node (if high availability is not critical at the initial stage) and scale the cluster later. For no-downtime migrations, PostgreSQL logical replication is typically used, as it allows replicating data between the old and new clusters without interrupting database availability. Please take a look at our available support packages, and we’ll be happy to assist you with the deployment, maintenance, and smooth migration of your database clusters: https://autobase.tech/docs/support |
Beta Was this translation helpful? Give feedback.
-
@vitabaks Sorry can't afford paid support but thanks for the offer anyway haha |
Beta Was this translation helpful? Give feedback.
-
@rausub Yeah I get it, it's possible to run with 2, it's not the point.. I was thinking about something like "reduced availability mode" where everything works and you can add more nodes as you grow... @vitabaks Why did you close this ? I thought this was a place for discussions..and now other users won't see it because closed discussions are hidden |
Beta Was this translation helpful? Give feedback.
-
Currently I can't use autobase due to it's hardware requirements (3 servers upfront is a high bar for young projects). But I will be closely following it and maybe use it later if/when my project can afford the hardware.
Will it be possible to migrate existing DBs to autobase without downtime (i.e. via replication) ? Are there any such plans ? Are there any ideas on how to allow small projects start with autobase earlier and to have autobase grow with them ?
Beta Was this translation helpful? Give feedback.
All reactions