#AIQUM capacity in Harvest

1 messages · Page 1 of 1 (latest)

native orchid
#

My AIQUM server has 13 ONTAP clusters and is configured as a data source for Harvest/NaBox, but only the first 4 clusters in the AIQUM database are being pulled. Shouldn't all the clusters in AIQUM be pulled by Harvest?
NaBox 2.5.1
Harvest 1.6

pearl salmon
#

Hi @native orchid I would think so. Harvest 1.6 has been out of support for quite awhile though and is no longer available. Perhaps you can find something in the log files

native orchid
#

There's nothing in the AIQUM_netapp-harvest.log about failure, just that it Discovered 4 clusters. Where else might I find reasons why the remaining 9 clusters aren't being 'discovered' by Harvest when polling UM? There are plans to upgrade both Harvest and NaBox, but looking to resolve this in the meantime.

pearl salmon
#

Not sure, never used that. Maybe @glossy mango knows?

glossy mango
#

Wasn’t there something used to match controllers with aiq entries ?

#

Like, when adding controllers or restarting the aiq poller needed to be restarted ?

#

All the clusters are healthy in aiq I suppose ? And you restarted everything already ?

native orchid
#

I've restarted/rebooted both servers, no joy.

glossy mango
#

All 13 clusters are configured in harvest ?

native orchid
#

No, there's a subset of the 13 in harvest, though it's only pulling 4 from UM. 3 are configured in harvest and 1 isn't. Oddly enough it's the 'first' 4 clusters in UM, as listed by ID in the UM database.

glossy mango
#

I thought I remembered harvest only pulled from UM what is configured in harvest

#

As an individual cluster

native orchid
#

[NORMAL ] WORKER STARTED [Version: 1.6] [Conf: netapp-harvest.conf] [Poller: UMSERVER]
[NORMAL ] [main] Poller will monitor a [OCUM] at [UMSERVER:443]
[NORMAL ] [main] Poller will use [password] authentication with username [] and password [**********] [NORMAL ] [sysinfo] Discovered [clusterA] on OCUM server and will submit metrics under group [EDC1]. [NORMAL ] [sysinfo] Discovered [clusterB] on OCUM server and will submit metrics under group [EDC1]. [WARNING] [sysinfo] Discovered [clusterC] on OCUM server but unable to submit metrics because no matching conf section found; to collect this cluster please add a section. [NORMAL ] [sysinfo] Discovered [clusterD] on OCUM server and will submit metrics under group [EDC2]. [NORMAL ] [main] Collection of system info from [UMSERVER] running [9.9P1] successful. [NORMAL ] [main] Found best-fit monitoring template (older generation or major release): [ocum-9.6.0.conf] [NORMAL ] [main] Added and/or merged monitoring template [/opt/netapp-harvest/template/default/ocum-9.6.0.conf] [NORMAL ] [main] Metrics for cluster [clusterA] will be submitted with graphite_root [netapp.capacity.EDC1.clusterA] [NORMAL ] [main] Metrics for cluster [clusterB] will be submitted with graphite_root [netapp.capacity.EDC1.clusterB] [NORMAL ] [main] Metrics for cluster [clusterD] will be submitted with graphite_root [netapp.capacity.EDC2.clusterD]

#

It discovers clustersA-D, but only sends A,B,D to graphite, which is fine. The issue is that it should 'discover' all 13 cluster

glossy mango
#

If my recollection is right, only if the cluster are configured as harvest source. But you’re saying it was working before ?

#

Mmm I see. You expect to see 10 times same error as C ?

native orchid
#

In essence, yes. Except that a few of the clusters NOT being discovered ARE in Harvest as data sources.

glossy mango
native orchid
#

It has been uploaded. Thank you for looking.

glossy mango
#

ok, so the gxxx cluster has no matching conf records so that makes sense but I still can't figure out why only 4 clusters are returned by AIQ

#

It might be worth updating the NMSDK package if not the latest already ?

native orchid
#

the NMSDK package is already at the latest.