Description
this bug manifest itself in the "show by server" page where we display curr_items for all the nodes.
the screenshot is captured from one of the users
the curr_items displayed in the UI toggles between 0 and 5 million.
the api call :
/pools/default/buckets/default/stats/curr_items?_=1308382785760
response where value was 0
{"tstampParam":1308382784595.0,"samplesCount":60,"isPersistent":true,"lastTStamp":1308382785595.0,"interval":1000,"timestamp":[1308382784595.0,1308382785595.0],"nodeStats":{"10.218.59.187:8091":[5390800,5390800],"10.76.58.246:8091":[5325256,5325269],"10.82.6.130:8091":[5306225,5306222],"10.82.21.98:8091":[5269926,0],"10.82.113.154:8091":[5222972,0],"10.82.113.173:8091":[4809062,4809070],"10.102.53.179:8091":[5299763,0],"10.117.26.161:8091":[5278740,5278760],"10.218.3.187:8091":[5289034,5289044],"10.218.37.191:8091":[5238413,5238418]}}
>>> [5299763,0],"10.117.26.161:8091
response where value was correct
{"tstampParam":1308382679595.0,"samplesCount":60,"isPersistent":true,"lastTStamp":1308382680595.0,"interval":1000,"timestamp":[1308382679595.0,1308382680595.0],"nodeStats":{"10.218.59.187:8091":[5390796,5390791],"10.76.58.246:8091":[5325221,5325232],"10.82.6.130:8091":[5306228,5306232],"10.82.21.98:8091":[5269914,5269916],"10.82.113.154:8091":[5222964,5222958],"10.82.113.173:8091":[4809050,4809047],"10.102.53.179:8091":[5299741,5299740],"10.117.26.161:8091":[5278743,5278728],"10.218.3.187:8091":[5289024,5289020],"10.218.37.191:8091":[5238397,5238398]}}
however if you switch to the "show all stats" for this node page the curr_items graph numbers don't toggle between 0 and 5 million anymore
it seems that ns_server does not wait long enough for getting stats from all nodes and fill in the value as zero for those nodes.