cassandra doesn't rotate data file
Bug #1392156 reported by
Jerry Zhao
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Compass |
In Progress
|
Undecided
|
Lawrence Davison |
Bug Description
cassandra doesn't have a retention policy for the moment and data will pile up until disk space is used up.
should set a retention policy.
current observation suggests that 90 nodes or less will write 32G bytes in 7 days.
Changed in compass: | |
assignee: | nobody → Lawrence Davison (lawrence-davison) |
Changed in compass: | |
status: | New → In Progress |
To post a comment you must log in.
While Cassandra may not, like most datastores, specify an explicit retention policy, we may be able to delete data based on a TTL or value proposition of certain collections. That said this is liekly to be very customer specific and will need to be configurable outside of the realm of a Chef Cookbook which is meant to be at least semi-ridgid per deployment. I suggest an alternate config system approach which is open to UI configuration.