I’m testing replication using docker compose with 2 containers running manticore each with its own data volume. My first problem is that manticore-extra package is not installed in 6.3.8 image, so I’ve built a custom image from manticoresearch/manticore:6.3.8 and added the extra package. After that replication works as expected. My table is replicated to the other node just fine.
Now the problem comes when I run docker compose down (but keeping the volumes). Then when I start the containers with docker compose up, what’s the proper procedure to recreate the cluster? Which case would apply from Manticore Search Manual: Creating a cluster > Setting up replication > Cluster recovery?
And since we are in containers manticore_new_cluster is not an option, nor adding --new-cluster to the searchd parameters.
For the moment I ran again in manticore-1 container:
CREATE CLUSTER qqq1 'qqq1' as path, 'manticore-1:9312,manticore-2:9312' as nodes;
alter cluster qqq1 add t;
And then in manticore-2:
JOIN CLUSTER qqq1 AT 'manticore-1:9312' 'qqq1' as path;
Finally I can query table “t” from both manticore containers. However, I’m not sure if it’s the proper workaround, especially when I have lots of tables.