toluaina
commented
1 year ago - First it is more efficient to run separate pgsync process per node
- So maybe separate each node out into its own schema.
- Does this still happen when you separate out the nodes into their own schema.
- It might also be easier the find the root issue causing the pending items.
mingchungchan
PGSync version: 2.3.3
Postgres version: PostgreSQL 13.2
Elasticsearch version: 7.17.0
Redis version: 5.0.6
Python version: 3.7
Problem Description: I have two nodes (A\ B) in the configuration of PGSync. At first I started PGSync and the data was synchronized. When the data change is heard in the subsequent monitoring, the first node A always stays in the waiting state of Redis, and the data cannot be synchronized to ES. The second node will not have such problems. Then I added the index name to the debug log printing, and found that after the node A was initialized and then obtain 1,000 data, it would not print the log anymore, and node B was still printing the log.
In addition, I thought it was too slow to query the data from Postgres. So I used the command to analyze the index. Then I added all the indexes. But this still does not work.
pgsync -c schema.json -aThanks.