Problems with dashboard.

[ibobak]

Andrew,

there is one more issue with the dashboard. Here is the stacktrace - see below. Feel free to ask me any data that you might need (e.g. to send you the backup of InfluxDB or anything like that). Just give me instructions how to give it to you.

Also feel free to give me any debug scripts to temporarily place instead of normal ones (to dump my data on the screen - I will paste it here).

Thanks.

GET /scripts/bootstrap.min.js 304 3.740 ms - - GET /css/style.css 200 430.211 ms - - /opt/influxdb-dashboard/public/scripts/influxdb.js:48 var series = seriesNames[0]; ^ TypeError: Cannot read property '0' of undefined at /opt/influxdb-dashboard/public/scripts/influxdb.js:48:26 at /opt/influxdb-dashboard/node_modules/influx/index.js:181:14 at /opt/influxdb-dashboard/node_modules/influx/index.js:64:14 at InfluxRequest._request (/opt/influxdb-dashboard/node_modules/influx/lib/InfluxRequest.js:97:12) at InfluxRequest._parseCallback (/opt/influxdb-dashboard/node_modules/influx/lib/InfluxRequest.js:115:19) at Request._callback (/opt/influxdb-dashboard/node_modules/influx/lib/InfluxRequest.js:107:10) at self.callback (/opt/influxdb-dashboard/node_modules/influx/node_modules/request/request.js:197:22) at Request.emit (events.js:107:17) at Request.onRequestError (/opt/influxdb-dashboard/node_modules/influx/node_modules/request/request.js:854:8) at ClientRequest.emit (events.js:107:17)

[ajsquared]

What version of InfluxDB and statsd-jvm-profiler are you using?

[ibobak]

profiler - 0.8.3 (made git clone 10 hours ago and built it). influxdb - 0.9.2-rc1 (installed using instruction from https://influxdb.com/download/index.html) OS = CentOS 6.6

the reason of using the last influxdb was the bug https://github.com/influxdb/influxdb/issues/3222 with stable 0.9.1, but frankly speaking in 0.9.2-rc1 it remains.

[ajsquared]

I've only tested the profiler with InfluxDB 0.9.1. Are you able to try that version?

If that doesn't work out, can you send me your dashboard-config.json file, the arguments given to the profiler when you used it, and the output of running SHOW SERIES on the database where the profiler writes its data?

[ibobak]

Andrew,

here https://drive.google.com/file/d/0B3DMXMfcPWF3SXo3ZG9hWGNEOTA/view is a document with detailed configuration and description of what I am doing and what I am getting. I am using the newest dashboard (just pulled it).

Please, help. Thank you in advance.

Regards, Ihor

[ajsquared]

The error in your original comment happens when the dashboard queries for the heap.total.max measurement. That metric was selected somewhat arbitrarily, but it should always be present and allows for a quick query to get all the valid tag values to populate those initial dropdowns.

It ultimately runs the query show series from "heap.total.max". Can you run that and show me the results?

[ibobak]

EDIT: here is what I get at work (with real mapreduce job). The original message I got at home. I will be there in an hour and will post an image what is inside there

[ibobak]

I am sorry if I confused you. I am little bit tired after 16 hours of work every day.

So, in general there are two things: 1) at home (with not a mapreduce job) I am getting problem "Cannot read property '0' of undefined" in the dashboard 2) at work (with a mapreduce job) I am getting the problem that the dashboard is empry (doesn't allow to select values for tags and build the Flame chart) - here is the screenshot: https://drive.google.com/file/d/0B3DMXMfcPWF3SXo3ZG9hWGNEOTA/view

[ajsquared]

Is this still a problem you're having after the changes you've made to deal with the OOM issues?

[ajsquared]

Let me know if you're still having any issues with this.