I am stuck on this issue. One of my customers cloned the production cluster (1 MA + 2 LEAVES) (MEMSQL 6.8) and started a new cluster on a different network out of it for testing purposes. Production and testing environments don’t communicate each other. Unfortunately the new cluster still has the production ip addresses. How can I change them?
Please refer to the attached picture where I am showing you the results of memsqlctl list-nodes, memsqlctl show-aggregators and memsqlctl show-leaves run against the new cluster. It’s still pointing to the old IP addresses
192.168.20.130 → MA,
192.168.20.133 → LEAF 1
192.168.20.134 → LEAF 2
I want to change them to the following IP addresses
The only command I am able to run is memsqlctl. Neither memsql-ops nor sdb-admin are available on the server.
Please can you describe the steps I should follow to change the IP addresses?
Here is some documentation related to your question. It requires sdb-admin though.
It might be simpler just to do a backup, script out the users & grants, restore the backup on a new cluster, then recreate the users & grants. You can export users & grants with sdb-admin dump
I managed to change the leaves IP addresses (via remove + add leaf commands). As you can see they point to the new IPs. But I cannot find a way to update the master aggregator’s IP. I want it to get 10.198.200.130 but it is still 192.168.20.130.
I add that running memsqlctl remove-aggregator --host 192.168.20.130 I get
Error 1749: You are attempting to remove the master aggregator
I don’t know what you could do. Maybe you could add a child aggregator, then promote it to master, then drop the old one, but I don’t know if that would succeed given the state the cluster is in. I’ll ask somebody else to take a look at this but don’t know for sure if there is a solution.
Apparently I can’t promote a node to master as soon as a master already exists. And if I try to remove the aggregator (that is the master) I get an error because I am trying to remove the mast aggregator.
