#making the switch from nabox to
1 messages · Page 1 of 1 (latest)
hi @upper loom what dashboard in Harvest 1.X had the nfs connection count that is missing in Harvest2. Maybe it got missed in the rewrite
I'll have to crank the old nabox server up again - gimme a minute
drwxr-xr-x 2 root root 4096 Nov 27 2020 .
drwxr-xr-x 11 root root 4096 Dec 8 2020 ..
-rwxr-xr-x 1 root root 8865 Aug 19 2019 bash_extension_example.sh
-rwxr-xr-x 1 root root 9407 Aug 19 2019 nfs-connections.sh
-rwxr-xr-x 1 root root 11529 Aug 19 2019 perl_extension_example.pl
-rwxr-xr-x 1 root root 11603 Aug 19 2019 python_extension_example.py
-rwxr-xr-x 1 root root 21880 Aug 19 2019 snapmirror_replications.py
nabox-api:/opt/packages/netapp-harvest/extension#
nfs-connections.sh
thanks @grizzled peak
ok, looks like that script ssh-es into cluster and issues ONTAP cli commands. That's not something Harvest 2 supports, but we can use the (preferred) REST private CLI to do the same
well fart - maybe it wasn't there - I can't find it now
@upper loom no need to crankup your old server - I see what's needed. If you have a GitHub account, would you mind creating a Harvest issue that captures this ask?
no worries
sure - np
thanks!
thanks! Rahul reminded me we added something like this in main recently 🎊
@upper loom Harvest has this information which is only available in our nightly builds . Template here https://github.com/NetApp/harvest/blob/main/conf/rest/9.12.0/netConnections.yaml
👍
This is only a Rest Collector feature. You would need to enable this template here https://github.com/NetApp/harvest/blob/main/conf/rest/default.yaml#L16 or add this line in custom.yaml. Add Rest collector to your pollers and you should see metrics like below
Do you mean how to enable this in NABox?
yes
Okay I'll list down the steps.
Thank you very much @wanton vine
Here are the steps https://github.com/NetApp/harvest/discussions/1284
@wanton vine thank you
is rest directory considered "standard" like perf and zapiperf ?
and is netConnection default with recent harvest builds ?
good ! Will get that in the main version
netConnections.yaml is only 9.12 ?
It uses private cli so should work in older versions also
ok so does it matter that's it's in the 9.12.0 directory and not 9.7.0 ?
it is not, harvest chooses based on information given here https://github.com/NetApp/harvest/blob/main/conf/README.md#harvest-versioned-templates
There may be need for different template for a version if counter name differs or it is absent
a custom.yaml file doesn't need collector: line, correct ?
object templates and harvest config are different. Collector: is only for harvest config and not for custom.yaml
thanks. DOes that seem ok to you ?
nabox-harvest2 | 2022-09-20T05:39:10Z INF rest/rest.go:184 > Using default timeout Poller=cluster2 collector=Rest:NetConnections timeout=30s
nabox-harvest2 | 2022-09-20T05:39:10Z INF collector/helpers.go:134 > best-fit template Poller=cluster2 collector=Rest:NetConnections path=/conf/conf/rest/9.12.0/netConnections.yaml v=9.7.0
nabox-harvest2 | 2022-09-20T05:39:10Z INF rest/rest.go:124 > initialized cache Poller=cluster2 collector=Rest:NetConnections numMetrics=0 timeout=30s
nabox-harvest2 | 2022-09-20T05:39:10Z INF collector/collector.go:344 > no instances, entering standby Poller=cluster2 collector=Rest:NetConnections task=data
nabox-harvest2 | 2022-09-20T05:39:28Z INF rest/rest.go:184 > Using default timeout Poller=AFF collector=Rest:NetConnections timeout=30s
nabox-harvest2 | 2022-09-20T05:39:29Z INF collector/helpers.go:134 > best-fit template Poller=AFF collector=Rest:NetConnections path=/conf/conf/rest/9.12.0/netConnections.yaml v=9.8.0
nabox-harvest2 | 2022-09-20T05:39:29Z INF rest/rest.go:124 > initialized cache Poller=AFF collector=Rest:NetConnections numMetrics=0 timeout=30s
nabox-harvest2 | 2022-09-20T05:40:01Z INF rest/rest.go:184 > Using default timeout Poller=cluster1 collector=Rest:NetConnections timeout=30s
nabox-harvest2 | 2022-09-20T05:40:22Z WRN ./poller.go:668 > init collector-object (Rest:NetConnections): 0: Poller=cluster1
nabox-harvest2 | 2022-09-20T05:44:10Z INF collector/collector.go:344 > no instances, entering standby Poller=cluster2 collector=Rest:NetConnections task=data
nabox-harvest2 | 2022-09-20T05:49:10Z INF collector/collector.go:344 > no instances, entering standby Poller=cluster2 collector=Rest:NetConnections task=data
no instance on cluster2 is fine I guess, that's a svm dr destination mostly
But I don't seem to find net_connection in prom
looks like no data on that cluster. Could you run network connections active show on that cluster and see?
cluster1::network connections active> show
Vserver Interface Remote
Name Name:Local Port Host:Port Protocol/Service
---------- ---------------------- ---------------------------- ----------------
Node: cluster1-01
vsphere lif1:2049 10.1.0.12:828 TCP/nfs
vsphere lif1:2049 10.1.0.11:1007 TCP/nfs
vsphere lif1:2049 10.1.0.11:1008 TCP/nfs
vsphere lif1:2049 10.1.0.12:829 TCP/nfs
docker lif1:2049 10.1.0.41:890 TCP/nfs
5 entries were displayed.
hmm strange . could you DM me the cluster details and i'll try locally
I am getting the data for the shared cluster
harvest version 22.09.19-nightly (commit 3d57787) (build date 2022-09-19T00:23:03-0400) linux/amd64
yeah that should be fine
Are you sure its the same cluster? shows 5 entries in cli for u as per screenshot above and for me i get 542 entries?
The other one I showed you was my home lab, this one you have access
ok do u mean u get no data for this cluster also?
indeed
wait I see that now :
nabox-harvest2 | 2022-09-20T06:07:31Z INF rest/rest.go:311 > Collected Poller=AFF apiD=28m2.239s collector=Rest:NetConnections instances=552 metrics=6624 parseD=8ms
cool
Lol, that's one of those things, I didn't wait enough.
Isn't that suppose to collect on the next minute ?
apiD shows 28m2.2s??
it took 28 minutes to get response?
it should collect in first poll itself
is that a single REST call or many ?
should be single in this case since records are tiny
only if it were many thousands, we go with batching
ok I'll test again on a local nabox