Make senses to me. Not sure if I should just push a Ninja fix, created the ticket anyway: CASSANDRA-14370. Are you interested in creating a quick patch for it? On Tuesday, April 3, 2018, 3:06:28 AM PDT, Nicolas Guyomar <nicolas.guyo...@gmail.com> wrote: Hi Jay,
Well the log in itself does not provide useful information (like segment number or sthg like that), so IMHO trace would be a better level for this one I agree that one log per sec may not be seen that verbose ! Thank you On 30 March 2018 at 06:36, Jay Zhuang <jay.zhu...@yahoo.com> wrote: > It's changed to trace() in cassandra-3.0 with CASSANDRA-10241: > https://github.com/pauloricardomg/cassandra/commit/ > 3ef1b18fa76dce7cd65b73977fc30e51301f3fed#diff- > d07279710c482983e537aed26df80400 > > In cassandra-3.11 (and trunk), it's changed back to debug() with > CASSANDRA-10202: > https://github.com/apache/cassandra/commit/e8907c16abcd84021a39cdaac79b60 > 9fcc64a43c#diff-85e13493c70723764c539dd222455979 > > The message is logged when a new commit-log is created, so it's not that > verbose from my point of view. But I'm also fine to change it back to trace. > > Here is a sample of debug.log while running cassandra-stress: > https://gist.githubusercontent.com/cooldoger/ > 12f507da9b41b232d8869bbcd2bfd02b/raw/241cd8f0639269966aa53e2b10cee6 > 13f8ed8cfe/gistfile1.txt > > > > On Thursday, March 29, 2018, 8:47:54 AM PDT, Nicolas Guyomar < > nicolas.guyo...@gmail.com> wrote: > > > Hi guys, > > I'm trying to understand the meaning of the following log > in org.apache.cassandra.db.commitlog.CommitLogSegmentManager.java > > logger.debug("No segments in reserve; creating a fresh one"); > > I feel like it could be remove, as it seems to be kind of a continuous task > of providing segment > > Any thought on removing this log ? (my debug.log is quite full of it) > > Thank you > > Nicolas >