Hello!
I have a cluster system running with 2 data nodes and 2 management nodes. I have a few tables that increase in size periodically. Over time I have noticed that the D8, D9, D10 and D11 DBLQH directories steadily increase in size. I understand that these are REDO log files and from what I've seen the only way to 'remove' these is to perform a rolling restart of the data nodes with the --initial flag (one at a time).
Is there a better way to ensure the REDO log files do not grow unbounded and fill the partition allocated to it? I am concerned that I do not have something configured correctly because I was under the assumption that the REDO log files would grow and shrink back down over time as checkpoints occurred across the cluster.
Thanks for any help!
Brian
I have a cluster system running with 2 data nodes and 2 management nodes. I have a few tables that increase in size periodically. Over time I have noticed that the D8, D9, D10 and D11 DBLQH directories steadily increase in size. I understand that these are REDO log files and from what I've seen the only way to 'remove' these is to perform a rolling restart of the data nodes with the --initial flag (one at a time).
Is there a better way to ensure the REDO log files do not grow unbounded and fill the partition allocated to it? I am concerned that I do not have something configured correctly because I was under the assumption that the REDO log files would grow and shrink back down over time as checkpoints occurred across the cluster.
Thanks for any help!
Brian