Enclsure C7000 server Mezzanine

[Ordos321]

Hello, currently in the configuration that I have we have that the SAN and LAN traffic goes through the 10Gb mezzanine connection with the Flex10 modules of the Enclosure. I would like to know if there is a way to measure the bandwidth used of those 10Gb in total as I think we have bandwidth problems and there are peaks where we may be saturating those 10Gb links.
Thank you very much
Regards.

[MGoorman]

What would you like to measure, the througput of the uplink of the Flex10 module, or the througput of your nic in your server?

The Flex10 has SNMP so you could check the ports with that.
Your server nic could also support SNMP or WMI (depending your OS).

[Ordos321]

Hi, I want to measure the bandwidth between the server and the Flex10 modules (the mezzanine). I would like to find out if the 10Gb link could fall short.

In my config I have 3 nic and 1 FcOE to connect to SAN in the OS side (Red Hat).

BR

[MGoorman]

I don't think that is possible, but probably the speed of your flexnics are variable.
Look at the defined speed in the server profile of the VirtualConnect..

It can be that your speed is 5-10G and that you share that speed with another flex nic.
That could be the reason that the speed is less than expected.

[support_s]

Hello,

Let us know if you were able to resolve the issue.

If you have no further query, and you are satisfied with the answer then kindly mark the topic as Solved so that it is helpful for all community members.

Please click on "Thumbs Up/Kudo" icon to give a "Kudo".

Thank you for being a HPE valuable community member.

[Ordos321]

Hello, the problem is that the bandwidth configured at the virtual connect level is greater than the 10Gb that the link has. If you add up all the NICs defined with their configured bandwidths, you get a theoretical maximum of more than 10Gb but I want to know if it is possible to measure the use of those 10Gb because one thing is the theory and another is the real use of the 10Gb link.

Best Regards

[MGoorman]

I still don't get what you try to achieve.

In my opinion its very simple:
You Flex-10/10D (assuming thats the Virtual Connect your using) has 16x 10Gb downlinks.
Im not sure how many blades you have provisioned, what blades you use and what NIC you use.

If you are using for example 2 blades with a 2-port 10Gb NIC.
And your provision 4 virtual NIC's to each blade, then you have in total 4 downlinks (1 for each 10Gb port, so 2 10Gb links per blade).
If you're using 4 virtual NIC's,. that the speed of all the NIC's should be variable from 5-10Gb (unless you manually edit the speed).
You simply get 20Gb to devide. So every virtual NIC should have at least 5 Gb.

Now; If I configure the same as you have (3x Ethernet Adapter, 1x FCoE HBA conn), then 1 of my Ethernet Adapters is 10Gb, the other 2 are 5-10Gb. 
Dont have a config for a fabric so I can't see speeds of the FCoE adapter.

These connections are made towards the FlexFabric/VirtualConnect, so connections between the blades should have at minimum that speed..
If you have some uplinks configured on the external ports, maybe those uplinks get congested what means you can't get the full speed between the blades and you external device.

What I should do; 
- Make sure the speeds of the virtual nics are what you excpect it to be
- Monitor your VirtualConnect with SNMP (at least the uplinks)
- Monitor your virtual NIC's on the blades

If you like, you can make more clear what issue you have, by providing the details/usecase etc.
Now its mostly guessing.

[MGoorman]

Also information from the quickspecs:
And maybe interesting: HP VC Cookbook

Performance
• 16 x 10Gb downlinks to server NICs
• Each 10Gb downlink supports up to 4 FlexNICs or 3 FlexNICs and 1 iSCSI FlexHBA
• Each iSCSI FlexHBA can be configured to transport Accelerated iSCSI protocol.
• Each FlexNIC and iSCSI FlexHBA is recognized by the server as a PCI-e physical function device with
adjustable speeds from 100Mb to 10Gb in 100Mb increments when connected to a HPE NC553i
10Gb 2-port FlexFabric Converged Network Adapter or any Flex-10 NIC and from 1Gb to 10Gb in
100Mb increments when connected to a NC551i Dual Port FlexFabric 10Gb Converged Network
Adapter or NC551m Dual Port FlexFabric 10Gb Converged Network Adapter including NC554FLB
Dual Port FlexFabric Adapter
• 4 x 10Gb cross connects for redundancy and stacking
• 10 x 10Gb SR, LR, or LRM fiber and copper SFP+ uplinks
• Supports up to 4 FlexNICs per 10Gb server connections.
• Each FlexNIC is recognized by the server as a PCI-e physical function device with customizable speeds
from 100Mb to 10Gb.
• Line Rate, full-duplex 600 Gbps bridging fabric
• 1.0 µs latency
• MTU up to 9216 Bytes - Jumbo Frames
• Supports up to 128K MAC addresses and 1K IGMP groups
• VLAN Tagging, Pass-Thru and Link Aggregation supported on all uplinks
• In tunneled VLAN mode, up to 4,096 networks are supported per network uplink and server downlink.
In mapped VLAN mode, up to 1,000 networks are supported on network uplinks per Share Uplink Set,
domain or module and on server downlinks up to 162 networks are supported per 10Gb physical port
(VC v3.30 or later).
• Stack multiple Virtual Connect Flex-10/10D modules with other VC Flex-10/10D, VC FlexFabric or VC
Flex-10 across up to 4 BladeSystem enclosures allowing any server Ethernet port to connect to any
Ethernet uplink

[Ordos321]

Hi, this is our current configuration:

So based on the configuration we have that 4 NIC + 1 FCOE go for each link of the 2 10Gb links it has. This makes a minimum of 2+2+2+2+4 and a maximum of 10+10+10+8, in order not to saturate the link none of the interfaces configured in the VC should pass their minimums at the same time but by configuration of the virtual connect you can have peaks where those 10Gb are exceeded.

At the uplinks level, we have configured 2 uplinks of 10Gb for the network part and 2 uplinks of 8Gb for the SAN part per module (have 2 modules). These uplinks are used by 5 blades.

Thanks

[MGoorman]

in order not to saturate the link none of the interfaces configured in the VC should pass their minimums at the same time but by configuration of the virtual connect you can have peaks where those 10Gb are exceeded.

Your calculation is right; the links have a minimum guaranteed bandwith.
If one link does not use any bandwith at all, another link can use that bandwith until the downlink reaches 10Gb.

If you like to set a maximum speed (minimum can be done by editing your profile > port speed > custom), you may be interested in reading this article 
https://support.hpe.com/hpesc/public/docDisplay?docLocale=en_US&docId=sf000057148en_us