SMB2/SMB3 performance data | NetApp | Page 1

short shuttle Oct 21, 2024, 4:35 PM

#

I seem to be missing data in all of the cards in this group. Is there a specific collector I should ensure is enabled?
These are the collectors I've got so far

#

I should say this is the current version I'm running:

keen hazel Oct 22, 2024, 6:22 AM

#

@short shuttle It is enabled by default https://github.com/NetApp/harvest/blob/main/conf/restperf/default.yaml#L50
https://github.com/NetApp/harvest/blob/main/conf/zapiperf/default.yaml#L57

Could you please share Harvest logs @ ng-harvest-files@netapp.com
https://netapp.github.io/harvest/24.08/help/log-collection/#nabox-4

keen hazel Oct 22, 2024, 11:52 AM

#

You can also upload support bundle here
https://upload.nabox.org/voha-koni-nuzi

short shuttle Oct 22, 2024, 3:13 PM

#

keen hazel You can also upload support bundle here https://upload.nabox.org/voha-koni-nuzi

Support bundle uploaded, if you'd also like me to grab the harvest logs let me know and I can grab those too

stoic wave Oct 22, 2024, 4:24 PM

#

Thanks for the support bundle @short shuttle I see from the log files that when Harvest tries to collect SMB2 information via RestPerf ONTAP returns a API request rejected => error making request StatusCode: 404, Message: Object \"smb2\" was not found., Code: 8585320 which means that table is not found on the ONTAP side. What version of ONTAP are your clusters? It looks like ONTAP added this table to RestPerf in 9.14.1+. Is it possible that you are using a version of ONTAP before then?

short shuttle Oct 22, 2024, 4:27 PM

#

stoic wave Thanks for the support bundle <@689858598968623117> I see from the log files tha...

All my systems are running this version:
NetApp Release 9.13.1P6: Tue Dec 05 16:06:25 UTC 2023

stoic wave Oct 22, 2024, 4:28 PM

#

ah ok, that would explain why you aren't collecting SMB2 via RestPerf. It doesn't exist on the ONTAP side. If you switch the collector to ZapiPerf, you should see the data collected

short shuttle Oct 22, 2024, 4:29 PM

#

would I lose the other data I'm collecting via REST? can I have both collectors running? ZAPI goes away at 9.14+ I think, right? Does that mean REST will have more collctors when ZAPI goes away?

stoic wave Oct 22, 2024, 4:31 PM

#

you won't lose anything. yes, both can be running at the same time. In fact, you already have both running at the same time. https://netapp.github.io/harvest/nightly/architecture/rest-strategy/#can-i-use-the-rest-and-zapi-collectors-at-the-same-time

short shuttle Oct 22, 2024, 4:34 PM

#

It looks like it might already be running?


collector:          ZapiPerf

# Order here matters!
schedule:
  - counter:  24h
  - instance: 10m
  - data:      1m

objects:
  # Node-level metrics
  CIFSNode:                 cifs_node.yaml
  Disk:                     disk.yaml
  ExtCacheObj:              ext_cache_obj.yaml
  FCVI:                     fcvi.yaml
  FcpPort:                  fcp.yaml
  FlexCache:                flexcache.yaml
  HeadroomAggr:             resource_headroom_aggr.yaml
  HeadroomCPU:              resource_headroom_cpu.yaml
  HostAdapter:              hostadapter.yaml
  Iwarp:                    iwarp.yaml
#  Netstat:                  netstat.yaml
  NFSv3Node:                nfsv3_node.yaml
  NFSv41Node:               nfsv4_1_node.yaml
  NFSv42Node:               nfsv4_2_node.yaml
  NFSv4Node:                nfsv4_node.yaml
  NVMfLif:                  nvmf_lif.yaml
  Namespace:                namespace.yaml
  NicCommon:                nic_common.yaml
#  ObjectStoreClient:        object_store_client_op.yaml
  Path:                     path.yaml
  Qtree:                    qtree.yaml
  Rwctx:                    rwctx.yaml
  SystemNode:               system_node.yaml
#  TokenManager:             token_manager.yaml
  VolumeNode:               volume_node.yaml
  WAFL:                     wafl.yaml
  WAFLAggr:                 wafl_hya_per_aggr.yaml
  WAFLSizer:                wafl_hya_sizer.yaml
#  NFSv4Pool:                nfsv4_pool.yaml

#

  CIFSvserver:              cifs_vserver.yaml
  CopyManager:              copy_manager.yaml
#  ExternalServiceOperation: external_service_operation.yaml
  FcpLif:                   fcp_lif.yaml
  ISCSI:                    iscsi_lif.yaml
  LIF:                      lif.yaml
  Lun:                      lun.yaml
  NFSv3:                    nfsv3.yaml
  NFSv41:                   nfsv4_1.yaml
  NFSv42:                   nfsv4_2.yaml
  NFSv4:                    nfsv4.yaml
#  NvmfRdmaPort:             nvmf_rdma_port.yaml
#  NvmfTcpPort:              nvmf_tcp_port.yaml
#  OntapS3SVM:               ontap_s3_svm.yaml
  SMB2:                     smb2.yaml
  Volume:                   volume.yaml
  VolumeSvm:                volume_svm.yaml
  WAFLCompBin:              wafl_comp_aggr_vol_bin.yaml
#  Vscan:                    vscan.yaml
  VscanSVM:                 vscan_svm.yaml

#  Uncomment to collect workload/QOS counters.
#  Workload:             workload.yaml
#  WorkloadVolume:       workload_volume.yaml

# The following workload templates may slow down data collection due to a high number of metrics.
#  WorkloadDetail:       workload_detail.yaml
#  WorkloadDetailVolume: workload_detail_volume.yaml```

stoic wave Oct 22, 2024, 4:36 PM

#

Nabox added this toggle, which I'm guessing you have on Auto? Try switching to Zapi or edit the underlying harvest.yml file that nabox is updating based on the toggle. For example, you have

collectors:
            - Rest
            - RestPerf
            - Zapi
            - ZapiPerf
            - Ems

if you change that to

collectors:
            - Rest
            - Zapi
            - ZapiPerf
            - RestPerf
            - Ems

you will collect config stuff via Rest and performance counters via ZapiPerf

#

Your current config has both ZapiPerf and RestPerf collectors specified. Since RestPerf is listed before ZapiPerf, RestPerf will be used first when an object exists in both collectors. That's outlined at the link I shared

#

you log files show that the collector being used is collector=RestPerf:SMB2

short shuttle Oct 22, 2024, 4:41 PM

#

Gotcha, but RestPerf doesn't have SMB2 according to the error? #1297961456326672447 message

I've swapped all my collectors (and the default so any new systems get this) to this:

            - Rest
            - Zapi
            - ZapiPerf
            - RestPerf
            - Ems```

#

now I need to restart Harvest I assume?

#

that would be this? dc restart havrest ? (I got this from another support email thread)

stoic wave Oct 22, 2024, 4:47 PM

#

Harvest has RestPerf templates for SMB2, but your cluster does not because ONTAP did not expose SMB2 counters to RestPerf until 9.14.1. Yes, please restart harvest with the dc restart command

short shuttle Oct 22, 2024, 4:56 PM

#

Thanks for that tidbit aroub REST vs ZAPI... my support team hasn't been able to explain what happens when ZAPI goes away. 🤣

#

I am now collecting data for those monitors! Thank you @stoic wave and @keen hazel!

#SMB2/SMB3 performance data