kjkoster
23-12-2010, 20:40
Dear All,
Just found an interesting presentation on monitoring Coherence. http://www.evidentsoftware.com/resources2/docs/2010-06-03_Coherence-Monitoring-Best-Practices-Evident.pdf It is an interesting read, even if you never deal with Coherence or cluster systems like Coherence.
They are quite concerned with GC pauses. Left unchecked, they may even cause nodes to be evicted from the Coherence cluster. That shows you how tight the timings are on distributed systems like these.
While that may be true for Coherence clusters, I feel that does not automatically carry over to other distributed systems. I have had clients worry about GC pause times when their systems are slow due to terrible SQL queries or badly tuned machines.
What I like about the article is that they urge you to make the problems visible before you act.
Kees Jan
Just found an interesting presentation on monitoring Coherence. http://www.evidentsoftware.com/resources2/docs/2010-06-03_Coherence-Monitoring-Best-Practices-Evident.pdf It is an interesting read, even if you never deal with Coherence or cluster systems like Coherence.
They are quite concerned with GC pauses. Left unchecked, they may even cause nodes to be evicted from the Coherence cluster. That shows you how tight the timings are on distributed systems like these.
While that may be true for Coherence clusters, I feel that does not automatically carry over to other distributed systems. I have had clients worry about GC pause times when their systems are slow due to terrible SQL queries or badly tuned machines.
What I like about the article is that they urge you to make the problems visible before you act.
Kees Jan