Red Hat Bugzilla – Bug 1017710
6 hour and 24 hour metrics expire too soon
Last modified: 2014-01-02 15:37:48 EST
+++ This bug was initially created as a clone of Bug #1017707 +++
Description of problem:
We use Cassandra's time to live (TTL) feature to let metric data age out automatically without having to explicitly delete it. Both 6 hour and 24 hour data are getting stored with the 1 hour data TTLs. Here are the prepared statements used to store the respective data,
insertSixHourData = storageSession.prepare("INSERT INTO " + MetricsTable.SIX_HOUR + "(schedule_id, time, " +
"type, value) VALUES (?, ?, ?, ?) USING TTL " + configuration.getOneHourTTL());
insertTwentyFourHourData = storageSession.prepare("INSERT INTO " + MetricsTable.TWENTY_FOUR_HOUR + "(schedule_id, " +
"time, type, value) VALUES (?, ?, ?, ?) USING TTL " + configuration.getOneHourTTL());
They are both using configuration.getOneHourTTL(), but it should be configuration.getSixHourTTL() and configuration.getTwentyFourHourTTL() respectively.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
Fix pushed to release/jon3.2.x branch.
commit hash: 9eac4656276f
Previous commit missed 6 hour ttl. I committed fix for this to release/jon3.2.x branch.
commit hash: 9a3eef65bb2ef6
Moving to ON_QA for testing in the next build.
Created attachment 822339 [details]
corresponding changes are committed to jon3.2.x branch - see screen-shot attached