2190900 series means you have an index with 2.2 million individual entries. That's going to require 16-32GB of RAM <https://docs.influxdata.com/influxdb/v0.13/guides/hardware_sizing/> just to hold the index. When you issue queries on that it requires more RAM for the query processing.
Try to lower your series cardinality by removing tags that have a high number of values, or writing them as fields instead. On Tue, Aug 16, 2016 at 6:39 PM, Himanshu Ghai <[email protected]> wrote: > > Also created https://github.com/influxdata/influxdb/issues/7160 > > Unfortunately I'd have to switch to cyanite/graphite if this didn't work > for me. Anything I could try at this time. > > > On Monday, August 15, 2016 at 6:21:13 PM UTC-7, Himanshu Ghai wrote: >> >> I suspect this query kills the db for a while >> >> show tag values from <measurement> with key=<keyName> >> >> Any help would be appreciated. My instance could not load a week's worth >> of metrics with only less than a gb of data on disk.. I must be >> doing something wrong here. >> >> Himanshu >> >> On Monday, August 15, 2016 at 5:57:40 PM UTC-7, Himanshu Ghai wrote: >>> >>> I tried upgrade to 1.0 beta but that did not help. Also tried to alter >>> default retention policy duration >>> "alter retention policy default on influx_prod duration 520w". it did >>> not help either >>> >>> i ran stats output as requested in the github issue. For the db i use it >>> has this value below >>> What does the numSeries value signify? >>> >>> name: database >>> tags: database=influx_prod >>> numMeasurements numSeries >>> --------------- --------- >>> 4 2190900 >>> >>> >>> >>> >>> >>> On Monday, August 15, 2016 at 12:24:41 PM UTC-7, Himanshu Ghai wrote: >>>> >>>> In this post https://github.com/influxdata/influxdb/issues/5440# >>>> issuecomment-233230997, >>>> >>>> jwilder commented to change the retention policy. Can I save my data >>>> and still follow >>>> these steps? The steps mention to create db and set retention policy. I >>>> did not change >>>> the retention policy and it is possibly running default. >>>> >>>> Thanks for your help >>>> Himanshu >>>> >>>> >>>> >>>> On Monday, August 15, 2016 at 11:53:48 AM UTC-7, Himanshu Ghai wrote: >>>>> >>>>> >>>>> >>>>> - My setup is running on a single 4 cpu instance on aws. 0.13 >>>>> version >>>>> - Ingests about 5000 measurements with 4 tags each every 3 hours. >>>>> - Disk has provisioned iops set to 2000 and i/o wait is not a >>>>> problem. >>>>> >>>>> >>>>> On simple select queries I get a lot of user cpu spikes(80-90%) and >>>>> timeouts. If I request again after a while, it may work fine. >>>>> >>>>> Can someone help me with pointers on what I should be looking at? >>>>> >>>>> Thanks for your help >>>>> Himanshu >>>>> >>>> -- > Remember to include the InfluxDB version number with all issue reports > --- > You received this message because you are subscribed to the Google Groups > "InfluxDB" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at https://groups.google.com/group/influxdb. > To view this discussion on the web visit https://groups.google.com/d/ > msgid/influxdb/fd2dd6c0-3c7e-478d-a43a-6df60ce9d3c6%40googlegroups.com > <https://groups.google.com/d/msgid/influxdb/fd2dd6c0-3c7e-478d-a43a-6df60ce9d3c6%40googlegroups.com?utm_medium=email&utm_source=footer> > . > > For more options, visit https://groups.google.com/d/optout. > -- Sean Beckett Director of Support and Professional Services InfluxDB -- Remember to include the InfluxDB version number with all issue reports --- You received this message because you are subscribed to the Google Groups "InfluxDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/influxdb. To view this discussion on the web visit https://groups.google.com/d/msgid/influxdb/CALGqCvPjYrsDdCZX3Mp0wTOWLgOdNefnUT0Fswuu8-4d0_sa_Q%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
