Details
-
Bug
-
Resolution: Fixed
-
Major
-
2.0-beta
-
Security Level: Public
-
None
-
centos 6.1 64 bit
Description
Install couchbase server 2.0.0c-704 on 4 nodes
10.3.121.23
10.3.121.24
10.3.121.25
10.3.121.26
Create default-s bucket and load 17 million items to that buckets
Create 11 dev views and 1 production view. Let the cluster indexing and query those views.
Let the load running about 3K ops
Set up auto failover 30 seconds to cluster
Shut down couchbase server on node 26 and autofailover kicked in
Rebalance to remove node 26. The first rebalance failed. The second tried passed
Let load running with set, get, mutate and expire.
I notice view disk size are so hug (above 20GB each) and compaction did not kick in.
I did manually click compaction button at view3. It did compact the view file size from 20+ GB down to 3 GB.
I monitor atop, see cpu are not high (about 70%) and disk always busy at 100%
Here are the links to diags, collect_info and atop files
https://s3.amazonaws.com/packages.couchbase/diag-logs/c-201208/3nodes-704-collect-info-compaction-20120818.tgz
https://s3.amazonaws.com/packages.couchbase/diag-logs/c-201208/3nodes-704-compaction-not-run-20120818.tgz
https://s3.amazonaws.com/packages.couchbase/diag-logs/c-201208/atop-3nodes-704-compaction-not-run-20120818.tgz