I am taking ttbackup on the virtual machine (RHEL 7.4 , ESXi, NL-SAS datastore) and this generates a huge IO and cache usage. This sometimes impacts the application which is running inside the VM (very latency intensive application).
Has anyone come across this? Is there any way to control the ttbackup in terms of cache usage or reduced IO?
There are no options or hooks for this in ttBackup itself. ttBackup essentially copies one of the database checkpoint files and a subset of the current set of transaction log files. It does this as quickly as it can and so it can certainly generate heavy I/O. You might be able to use OS level tools to throttle ttBackup (maybe renice it to lower its priority) and perhaps also reduce the maximum size allowed for the OS cache.