Details
-
Bug
-
Resolution: Fixed
-
Critical
-
6.5.0
-
None
-
Untriaged
-
Centos 64-bit
-
-
Yes
Description
Steps are here:
1. load data to cluster until you hit 85% resident ratio - basically get the cluster to DGM, cluster has 3 buckets, with replicas
2. exp_pager_stime to 10 on 3 buckets on all nodes
3. Validate the stats for active and replica vbuckets
Validation of active and replica items is failing for stats.
All warmup tests in regression fails during verification of stats after loading data on a 4 node cluster.
curr_item_total:
Expected : 2460000
Actual: 2457767
vb_replica_curr_items:
Expected: 1230000
Actual : 1227767.
whereas vb_active_curr_item is 1230000.
Attached memchached failure and the test failure logs snippet. All gave links for all the cbcollect logs
Link to testrunner test command:
./testrunner -i b/resources/4-nodes-template.ini -p max-dupe-result-count=10,num-tries=60,attempt-num=60,dgm_run=true,eviction_policy=fullEviction,is_container=True -t warmup.warmuptest.WarmUpTests.warmup_test,stats_monitor=vb_replica_perc_mem_resident;vb_active_perc_mem_resident,items=20000,sasl_buckets=1,standard_buckets=1,active_resident_threshold=90,doc_ops=update,GROUP=P0
Attachments
Issue Links
- mentioned in
-
Page Loading...