Thanks for you answer, this is what I was looking for. https://ignite.apache.org/releases/latest/javadoc/org/apache/ignite/cache/affinity/rendezvous/ClusterNodeAttributeAffinityBackupFilter.html ________________________________ De : Jeremy McMillan <j...@gridgain.com> Envoyé : jeudi 30 mai 2024 19:33 À : user@ignite.apache.org <user@ignite.apache.org> Objet : Re: Best way to update and organize nodes
This could work if you set up availability zones and use backup filters. Then you could perform maintenance one entire AZ at a time. When running during maintenance, your workload might exceed the capacity of the fraction of server nodes remaining up, so beware that. On Thu, May 30, 2024, 11:30 Louis C <l...@outlook.fr<mailto:l...@outlook.fr>> wrote: Hello everyone, I had a question that I could not really answer reading the documentation : Let's say I have a cluster of 10 Ignite server nodes, with one cache with persistent data and 2 data backups. I want to update the different nodes while maintaining the cluster activity (answering the clients requests). To do so I can stop gracefully one node, update it, and restart it, and then take care of the following nodes in the same fashion. In my understanding, this should ensure that no data is lost and that the cluster is still active (is this really the case ?). But this is quite long. I wanted to know if it was possible to set the different partitions in such a way that we know that we can shutdown half (or 1/3) of the nodes in the same time, to speed up this process. I guess it would be as if we have 5 primary nodes and 5 backups nodes, and that the 5 backup nodes take over when the 5 primary nodes shut down. Is such a thing possible ? Best regards, Louis C.