Disk usage is growing day-by-day : Couchbase Community edition 2.1.1

In production we have the following setup…
1) 3 couch base servers in a cluster
2) We have a Data Bucket "retailmall-session"
3) Each node for “retailmall-session” has bucket type (couchbase), RAM (2048 MB [total 6144 MB (2048 MB x 3 nodes)]), Replica Set (2), Reader/Writer (5), Compaction set to 30%.

So far we have Item Count as (218850) with RAM/Quota Usage (239MB /6GB), but the Data/Disk Usage is showing 22GB /27.3GB. We have set the expiration time as 30 minutes for each item we create/modify.

Why is that disk usage keep on growing day-by-day? We know that our http session object we store per item is no more than 3-6kb.

Why is the disk size growing ?


"Before 2.2, Couchbase Server storage engine would keep all deleted items on disk as “tombstone”, which is required by XDCR and other clients to recognize deleted items from brand new items and thus enforce correct conflict resolution algorithm during replication. The major downside of storing the tombstones is that server may end up with unbounded storage space and significantly slow down disk operations for the server.  Tombstone purger remove old deletion records after deletion has been persisted for a given amount of time. "

source -https://docs.google.com/document/d/1Mh6VuZVkmvxg567h0fDLGOJNzQ4O5XSde1MzUuZ42LI/edit?pli=1

So what should be done for Couchbase community edition 2.1.1 version ? How should we clean up deleted items ?

Try this.

curl -u Administrator:password -X POST http://localhost:8091/pools/default/buckets/YOUR_BUCKET_NAME/controller/unsafePurgeBucket

–Source: https://groups.google.com/forum/#!topic/couchbase-8091/coADcepVnV4

Or better yet upgrade to Couchbase 2.2.0 CE it has meta data (tombstone purge) integrated with compaction.