Monday, 2014-09-29

*** openstack has joined #magnetodb12:41
isviridovaostapenko, sorry was on meeting13:15
isviridovaostapenko, what about POST?13:15
*** charlesw has joined #magnetodb13:28
aostapenkoisviridov, I thought that POST will be convenient for future adding some options, but now I consider GET as a better option13:42
isviridovGET is more convenient when you use browser13:43
isviridovI prefer v1/healthcheck13:44
aostapenkoisviridov: agree13:46
isviridovI see your update. Let us make ERROR uppercase13:50
aostapenkoisviridov, agree. Done13:57
*** jeromatron has joined #magnetodb14:33
*** k4n0 has quit IRC14:34
*** jeromatron has quit IRC14:43
*** jeromatron has joined #magnetodb14:44
*** ajayaa has quit IRC14:58
ominakovjeromatron, hello15:16
jeromatronHi ominakov15:16
ominakovjeromatron, i have one question about JMX metrics in C*15:17
ominakovshould we send requests to all nodes in cluster if we want to know such metric as "estimateKeys" in CF?15:18
jeromatronI believe that's per node yes.15:25
ominakovand we should divide sum of this numbers on replication factor?15:29
openstackgerritIllia Khudoshyn proposed a change to stackforge/magnetodb: Add 'last_updated' to 'table_info'
*** jeromatron has quit IRC15:39
romainhominakov: estimateKeys() is the sum of estimated keys in each SSTable of the node. So, yes, you have to invoke it on each node if you want cluster wide metrics.15:54
ominakovromainh, thx15:55
romainhominakov: estimation is based on index size, so you have a estimation of all keys, regardless replication factor16:01
romainhdepending what you want, you may want to divide by replication factor16:02
*** vivekd has quit IRC16:19
*** rushiagr is now known as rushiagr_away16:31
*** rushiagr_away is now known as rushiagr17:09
*** romainh has left #magnetodb17:37
*** rushiagr is now known as rushiagr_away18:42
openstackgerritIllia Khudoshyn proposed a change to stackforge/magnetodb: Add 'last_updated' to 'table_info'
*** charlesw has quit IRC22:26

Generated by 2.14.0 by Marius Gedminas - find it at!