Why? If we keep only a subset of logs do we need to bother compressing? Will it impact splunk? Today we keep 50 server.log and it's rotated based on size. I think this comes out of #108 which was initially thought to be a cause of running out of disk. I saw it was heap dumps in the bin dir that was really using disk on at least one node (6.5GB) vs logs (<500MB).
Why? If we keep only a subset of logs do we need to bother compressing? Will it impact splunk? Today we keep 50 server.log and it's rotated based on size. I think this comes out of #108 which was initially thought to be a cause of running out of disk. I saw it was heap dumps in the bin dir that was really using disk on at least one node (6.5GB) vs logs (<500MB).