Compaction and High CPU & RAM utilization
I'm evaluating performance of Couchbase for file storage and I started my evaluation of 2.0.0 on single node with 4GB RAM and 100GB HDD.
I wrote a PHP script to push and retrieve files, read and write time on 2.0.0 implementation are impressive. Out of various issues faced during evaluation I faced regular server crashes because of high CPU utilization (http://www.couchbase.com/forums/thread/occasional-high-cpu-usage-beam-sm...).
Because of above issue I upgraded to 2.0.1 on 3 node cluster with 4GB RAM and 10GB HDD. After modifying my earlier PHP script, I started evaluating write performance.
The issues faced on 2.0.1 implementation are:
1. Frequent crashes because of High CPU and RAM utilization
2. Regular process failure to write a batch of 100 small files.
After checking log I can see compaction daemon is running every 30 secs, in spite of it being configured to be executed between 6 to 8 AM only (http://www.couchbase.com/forums/thread/compaction-daemon-not-following-a...).
In order to verify, if there is something wrong with Couchbase implementation I bumped up another 6GB of per node. After this the issue of crashes because of High CPU and RAM utilization is minimal, but regular process failure is still there.
Further to verify, I removed 1 node from cluster and ran same PHP script to write file in Couchbase and I'm still facing regular process failure on single node.
I'm wondering what went wrong in the upgrade?
Thanks for your help.
No answers yet