Harvest is sending performance data to graphite but stops sending capacity data

[matthias_beck]

Hi guys,

Versions:

Harvest: 1.6

Graphite: python-carbon-0.9.16-1.el7.noarch

Graphite and Harvest are installed on different servers.

OS: CentOS Linux release 7.8.2003 (Core)

After my vacation I recognized, that Harvest is not sending the capacity information anymore to graphite and I have no idea why. If I restart the graphite server and stop and start Harvest it starts sending the data, but stops in between and after some time it stops completely.

I checked the logs and could not find anything. In one of the poller logs I saw the following warning: 

[2020-09-14 14:41:50] [WARNING] [volume] data-list poller refresh overdue; skipped [1] poll(s) from [2020-09-14 14:42:00] to [2020-09-14 14:42:00]
[2020-09-14 14:41:50] [DEBUG ] [volume] data-list poller next refresh at [2020-09-14 14:42:00]

Any I idea where I have to look for a better analyze?

Thx.

Regards

Matthias

[ttran]

Hi Matthias_beck,

Can you try upgrading Harvest to 1.6.1 as it contains fixes and enhancements for both Graphite and specifically capacity statistics?

More information and upgrade installation for Harvest 1.6.1 are located here:

Harvest 1.6.1 

Regards,

Team NetApp

[bkamil]

Capacity stats are usually collected from UM, not directly from the clusters. Check the logs for UM workers of Harvest and make sure there is no connectivity and/or auth issues between Harvest and UM.