I've noticed that almost all RCLI scripts (ESX 4 or 3.5) take anywhere between 5-30 seconds to complete on the vMA. But when running directly on the console, the output is returned instantaneously. Has anyone had this problem? 5 seconds seems like its not such a big deal, but in reality, it can add hours to an automated test suite that uses the vMA to communicate to all ESX hosts.
[ibm3550-10]# time esxcfg-scsidevs -a
Adapter_ID Driver UID PCI Vendor & Model
vmhba0 aacraid pscsi.vmhba0 (2:0.0) ServeRAID 8k/8k-l8
vmhba1 qla2xxx fc.2000001b3200d037:2100001b3200d037 (7:0.0) ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba2 qla2xxx fc.2001001b3220d037:2101001b3220d037 (7:0.1) ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba3 ata_piix ide.vmhba3 (0:31.1) 631xESB/632xESB IDE Controller
vmhba32 ata_piix unknown.vmhba32 (0:31.1) 631xESB/632xESB IDE Controller
vmhba33 iscsi_vmk iqn.1998-01.com.vmware:ibm3550-10-39acade4 () iSCSI Software Adapter
real 0m7.522s
user 0m3.760s
sys 0m0.208s
Same command on the console:
vmhba0 aacraid link-n/a pscsi.vmhba0 (2:0.0) Adaptec ServeRAID 8k/8k-l8
vmhba1 qla2xxx link-up fc.2000001b3200d037:2100001b3200d037 (7:0.0) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba2 qla2xxx link-up fc.2001001b3220d037:2101001b3220d037 (7:0.1) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba3 ata_piix link-n/a ide.vmhba3 (0:31.1) Intel Corporation 631xESB/632xESB IDE Controller
vmhba32 ata_piix link-n/a ide.vmhba32 (0:31.1) Intel Corporation 631xESB/632xESB IDE Controller
vmhba33 iscsi_vmk link-n/a iqn.1998-01.com.vmware:ibm3550-10-39acade4() Software iSCSI
real 0m0.024s
user 0m0.000s
sys 0m0.020s
The protocol used is https@443. I've tried changing it to http@80 but this caused the following cryptic syntax error in one of the RCLI perl libs.
[ibm3550-rtp10]# time esxcfg-scsidevs -a --protocol http --portnumber 80
Can't call method "getChildrenByTagName" on an undefined value at /usr/lib/perl5/5.8.8/VMware/VICommon.pm line 1884.
real 0m1.685s
user 0m1.517s
sys 0m0.161s
i suspect the slowness is due to the authentication piece, has anyone figured out why?
Rob
Hi Rob, the slowness is due to vCLI (RCLI) itsself, instead of vMA fastpass authentication. And you will see performance improvement in the next vMA/vCLI release.
Lets hope that is the case and we'll see an update soon and not have to wait for a major release of ESX(i) to get a new vMA
=========================================================================
William Lam
VMware vExpert 2009
VMware ESX/ESXi scripts and resources at:
VMware Code Central - Scripts/Sample code for Developers and Administrators
If you find this information useful, please award points for "correct" or "helpful".
One can only hope
ty
-rob
Do you have a date when the next version of the RCLI is coming out? The performance is painful. Also any idea why only https works and not http? I want to understand whether its a configuration issue on my end
esxcfg-scsidevs -l -protocol http
Error: Server version unavailable at 'http://dvsimenane06.sim.nane.netapp.com:443/sdk/vimService.wsdl'
Some updates regarding performance of vMA?
Becouse its quite a reason for not using vMA.