4 ms·
Our ElasticSearch cluster was running out of RAM. We were doing everything in 32-bit mode, for performance (we couldn’t do Compressed OOPs for some reason) so e
by krallja 3y ago
Our ElasticSearch cluster was running out of RAM. We were doing everything in 32-bit mode, for performance (we couldn’t do Compressed OOPs for some reason) so each node could handle, effectively, 4GiB of data. And we didn’t have the budget for a new machine.
But I did notice that the most-recently-added node had 16GiB of physical RAM, for some reason. So I added a hypervisor and spawned two more VMs inside of it, each running ElasticSearch as well, thus increasing our cluster capacity by more than one year of growth!