#Harvest - switch from Zapi to Rest

1 messages · Page 1 of 1 (latest)

radiant current
#

Hello, what obvious I'm doing wrong with my harvest configuration?

We run ONTAP version 9.15.1P6 with Harvest version 25.02.0-1. The setup has been working with Zapi collector for several years but now I've been trying to move to Rest and I'm missing something obvious but can't figure out what.

When I either switch the order for the collectors from Zapi to Rest or just comment out Zapi in my harvest.yml file the polling stops working.

The log gets filled with "Skip instanceKey, not found in cache Poller=xxxxxx collector=RestPerf:*"

And earlier when selecting the template, which I think is the problem somehow: "best-fit template Poller=xxxx collector=Rest:SecuritySsh path=conf/rest/9.12.0/security_ssh.yaml v=9.15.1
init collector-object error="implementation error => PollCounter" Poller=xxxx collector=Rest object=SecuritySsh"

Ideas/suggestions?

compact cypress
#

Have you made any changes to default.yaml file under conf/rest?

radiant current
#

Logs uploaded, no changes done. Just made sure by replacing the file from source

compact cypress
#

From the Rest Logs, I see that harvest version used is harvest version 23.11.0-1 (commit 166b0263)
Could you upgrade to latest version? 25.02

#

Please share logs from latest Harvest 25.02. You can check used Harvest version from first log line post start Harvest

2025-03-27T07:13:29Z INF poller/poller.go:215 > Init Poller=xxxx asup=false confPath=conf config=/opt/harvest.yml configPath=/opt/harvest.yml daemon=false debug=false homePath= logLevel=info logPath=/var/log/harvest/ profiling=0 promPort=13000 version="harvest version 23.11.0-1 (commit 166b0263) (build date 2023-11-13T14:57:09+0000) linux/amd64"

radiant current
#

🤦‍♂️ thank you, works now. I'll need to check why my command for creating the docker container had gotten stuck to that old harvest version. No idea how I've been missing it.

compact cypress
#

I have noticed an error from your Zapi logs as below

2025-03-27T07:15:37Z ERR zapiperf/zapiperf.go:1486 > add instance error="duplicate instance key => xxxx" Poller=amscfs01 collector=ZapiPerf:FlexCache

Do you still see it with 25.02?

radiant current
#

No I don't. The only error on the logs is for the admin node connection which is expected.

#

This might be an entirely another thread but one of the key reasons why we wanted to switch to rest was to get FSA working in Grafana. It now dows, however there seems to be some limitation on Grafana that it only shows the top level qrtee directories in a volume.

e.g. on the NetApp GUI:

volume-AAA -> File System tab -> Explorer -> Directory names:

user123 (qtree)

projectA (qtree2)
projectB (qtree3)
projectC(qtree4)
user124
projectA(qtree2)
projectB(qtree3)
projectX(qtreeX)
user125
user126
etc.

The GUI allows us to see a breakdown of which projects are the top level qtrees (users) using most of their quota for. In Grafana there is no drilldown from the top level qtree in the volume in question. Not sure if my explanation makes sense and which might be the right channel to ask about this or make a suggestion to get this added as the functionality is already in the GUI so maybe not impossible to make available in Grafana side either.

compact cypress
#

You are in the right channel for Harvest related feature requests.

radiant current
#

Thanks, unfortunate but certainly makes sense atleast if the performance hit would be anything noticeable.

#

Not sure if feasible or doable but in this type of quota/disk usage case it would be enough to do the query daily or even weekly. Would not need to be realtime.

compact cypress
compact cypress
radiant current