Hi,

First of all, thanks for this fling, it's very useful and user friendly !

I used version 2.6.0 using easymode with FIO and looking at the graphs, I can see that each run as a duration of 1h30m.
If I understood well the FIO conf, we have 30 minutes of ramp up, and then 1h of benchmarch with logging. Looking at the FIO manual, it states that rampup_time is : "fio will run the specified workload for this amount of time before logging any performance numbers."

What I don't understand then is why latencies are increasing a lot when the rampup_time is over. If it's just logging of the performance numbers, it shouldn't have a huge impact on the bench ?

Thanks in advance for your input

Tristan

fio doesn't report the latency when doing the ramp up, it would only report that during the testing phase.

I have used HCIbench successfully in the past, but am having a problem this time. Our network is restricted to only IPV4 but my HCIbench deployment fails. In the logs I see this:

Not able to ping VMs ["hci-vdb-datastore-92-0-1"], try another time...
PING fe80::250:56ff:feb1:6c34(fe80::250:56ff:feb1:6c34) 56 data bytes
From fe80::250:56ff:feb1:caa4%eth0: icmp_seq=1 Destination unreachable: Address unreachable
From fe80::250:56ff:feb1:caa4%eth0: icmp_seq=2 Destination unreachable: Address unreachable
From fe80::250:56ff:feb1:caa4%eth0: icmp_seq=3 Destination unreachable: Address unreachable
From fe80::250:56ff:feb1:caa4%eth0: icmp_seq=4 Destination unreachable: Address unreachable
From fe80::250:56ff:feb1:caa4%eth0: icmp_seq=5 Destination unreachable: Address unreachable

--- fe80::250:56ff:feb1:6c34 ping statistics ---
5 packets transmitted, 0 received, +5 errors, 100% packet loss, time 48ms
pipe 4
Not able to ping VMs ["hci-vdb-datastore-92-0-1"], try another time...
Can't Ping VMs ["hci-vdb-datastore-92-0-1"] by their IPs ["fe80::250:56ff:feb1:6c34"]

I don't see any choice in HCIbench setup to choose IPV4 or IPV6, but it seems it now is expecting IPV6 to properly switch in our network. How do I configure HCIbench to only use IPV4?

Thank you,

Tim

this seems like link-local ipv6, meaning your vms did not get ipv4 addr. check your dhcp server.

I ran HCI bench on VMC and the test said it completed but there are no results generated. The .txt and .xls file are created, but no results are present. I tried all of the connectivity tests and they succeeded.

Hi Zack,

we recently noticed that HCIBench might not work properly in VMC due to some issues. would you mind try out this upcoming version instead and lets us know if this works or not? if not, we may want to shoot for a meeting to get that fixed.
http://ftpsite.vmware.com/download/1624568412-HCIBench_2.6.0.ova

Thanks
Chen

I installed 2.6.0, but I can't log in. The root password works on the console just fine, but not on the web login at 8443.

are you using the one on the ftp? could you try the version we post here?

I redeployed the version from this flings page and it worked. I ran a test and the reports were generated. The Easy Run does not work though. It just sits there and never deploys any workers. I have a one host VMC SDDC if that makes any difference.

would you mind running easy run and sending all the logs in /opt/automation/logs to vsanperformance@vmware.com?

Hi.

I just deployed the HCIBench ova and, even I have configured the network information (IPv4, DNS, mask, gateway), when I look at the Guest OS information in the vCenter console, I just can see two IPv6 addresses...no trace of the IPv4 address configured during deployment.

Any thoughts?

Thanks!

did you choose to use static when doing deployment?

Hi Chen Wei.

Yes, I choose Static but, even though that was right, you made me look at the configuration twice.

My mistake was a typo in the Net Mask. I wrote 255.255.255.128.0 instead of 255.255.255.128.

Corrected the typo, and the deployment and the connection were just right.

Thanks and best regards.