Thanks for reply.
But please check the below images, seems there is a problem on VC module:
Also I rebooted the server and put it to "Inteligence Provisioning" mode and it can detect first NIC.
↧
Re: HP BL460c - lost network connection on FlexFabric - Bay1
↧
Re: HP BL460c - lost network connection on FlexFabric - Bay1
Hi,
Since you have tried with a different enclosure, with a different blade and a different VC module, all seems to indicate the issue is located at the FlexFabric module.
Have you verified from the VCM GUI console if you have any Warning? Maybe the port has been "administratively disabled" becasue a pause flood problem? In that case:
o Click Re-enable Ports in the VCM GUI. or
o Use the "reset port-protect" CLI command.
If that still doesn't work, please open a formal support case to evaluate a HW replacement.
↧
↧
Re: Pause Frame Flood problem question
For what it's worth, we saw a similar issue on Integrity systems with that same Broadcom device and a Procurve switch.
We were told it was fixed with a new Broadcom driver last fall (2014). It's not clear to me if it was a run-time driver or a boot-driver (UEFI). The problem report does not show the new driver version, unfortunately.
Good luck,
Sue L
p.s. I am an HP employee.
↧
Re: Virtual Connect Command to Enable/Disable Downlink Port
That is exactly what I was looking for. I had seen the Speedtype=disable for uplinks but was hoping that option was available for downlinks. Thanks!!
↧
Re: HP BL460c - lost network connection on FlexFabric - Bay1
The down link is up now, there was no pause flood problem. The server restarted and downlink is in linked status.
But the link doesn't pass traffic.
I'm sure there is no misconfiguration on switches and virtual connect, also the Flex-Fabric module was working with our G7 server, but the module has problem on G8 server from first installation till now.
Do you think that upgrading VC firmware can resolve the issue? Or do you have any idea?
I can replace the module with another one easily, all other servers are production.
↧
↧
Re: 2012 R2 Drivers for NC542M 10GB Mezz Adapters
Hello
I would first recommend contacting HP Total Care for a workaround for drivers, but since you are out of warranty, I agree with
↧
Re: HP BL460c - lost network connection on FlexFabric - Bay1
Hi,
Yes, both test (VC fw upgrade and VC replacement) are worth trying.
Let us know the results.
David
↧
Virtual Connect (VC) communication to another enclosure
A question from Sanyi:
***********
Can I connect two VC FlexFabric module in different C7000 chassis directly? I don’t want to put both c7000 in the same domain I just want to connect them without a switch in order to both C7000 could communicate directly with each other.
***********
Reply from Robert:
************
You can use this knowledgebase article. It should help:
http://h41302.www4.hp.com/km/saw/view.do?docId=mmr_kc-0110609&hsid=64648787&sz=9827
***********
From Dan:
***********
Yes but the VLAN/Network(s) you plumb can only be access by blades in both chassis.
You cannot bounce that network out another uplink and up to the customer’s core due to the way VC handles networks.
So vMotion, SQL Heartbeat, Oracle RAC, etc are all fine as long as all the members are blades.
When you setup the link between the 2 sides, you need to change the mode from Auto to Manual/Failover or whatever it says.
This way you can control which link they use to talk and which link is your backup.
************
Other comments?
↧
Device bays show status "other"
We have a C7000 chassis that has been running with no issues for a while. Over the weekend, the active OA module locked up. It didn't failover, it just didn't respond. I physically removed and reinserted the OA and the secondary one took over and everything seemed OK
When I logged into Virtual Connect, I got a popup that I needed to re-enter the OA credentials (I don't remember the exact wording). I re-entered them and after a little while Virtual Connect came up fine.
However, in the OA console, every blade now has the while circle with the exclamation mark and shows a status of "other". Everything seems to be working, but how do I clear those status? Do I need to remove/readd each blade to the chassis?
Can/should I reset the VC Manager? Will that cause any outage to the blades?
This is a production environment, so I don't want to do anything to interrupte operations of the blades.
↧
↧
incompatible FC 4GB module after firmware upgrade
I have upgraded Firmwares on my c7000 enclosure, OA firmware upgrade went fine but VC is having message incompateble.
resetting did not bring any change.
I have even reomved all the blades from enclosure.
here are the current versions.
any suggestion how i can fix this incompatible error.
HP 1/10Gb VC-Enet Module On 3.61 2014-02-14T21:18:58Z
Bay 2 (LAN) OK HP 1/10Gb VC-Enet Module On 3.61 2014-02-14T21:18:58Z
Bay 3 (SAN) Incompatible HP Virtual Connect 4Gb FC Module On 1.30 7.12.0.14
Bay 4 (SAN) Incompatible HP Virtual Connect 4Gb FC Module On 1.30 7.12.0.14
↧
Flow Control question for Virtual Connect (VC)
Laura had customer question on Flow Control settings in VC:
*************
I have a customer that has c7000 enclosures, each with four VC Flex10-10D modules – two are leveraged for NetApp connectivity. NetApp and VMware are recommending that Flow Control be disabled. Because HP enables Flow Control by default & our documentation indicates this is standard practice, they are concerned about the ramifications of disabling it.
Can anyone tell me what HP’s recommended Flow Control setting in this scenario?
***********
From Brian:
I would go with the storage vendors recommendation typically. For example StoreVirtual recommends flow control be enabled….
And input from Pedrag:
I doubt you ll find recommended setup on this. But what I can share is my latest experience with flow control where we lost all traffic within c7000. We hit into buffer limit due to enabled flow control on downlinks. And the only recommendation was to disable flow control on downlinks. As soon as we done that, we got traffics and rx dropped packets gone. Flow control is OS dependent mechanizm, so you should be careful. Also, it should be followed by firmware/driver compatibility guides. There is also advisory regarding the flow control and pause frames.
Info from Kelly:
I worked with a customer who had pre-4.xx VC firmware, we hit some pause frame flood issues (enclosure outages), and due to schedule and timing, we tried as a gap-fix to turn off flow control for a couple weeks until we could get a maintenance window. We were initially advised by support the command (to change flow control) was non-disruptive. After we issued the command (live), we had multiple ping drops and lost one Windows server (had to e-fuse it). After – I recall we determined all links are dropped and re-negotiated with the change. Someone may need to confirm that….
Also - I recall a difference of where flow control is enabled by default, I recall learning the downstream ports are on, but the upstream (uplinks) are disabled by default.... someone may need to confirm that too. Via CLI, it does have THREE settings: auto|on|off
->set advanced-networking -help
Description:
This command sets the Ethernet advanced network tunable parameters.
Any changes may cause a network outage.
Syntax:
set advanced-networking [-quiet] [PacketBufferOverallocationRatio=<ratio>]
[FlowControl=<auto|on|off>]
Options:
quiet : This option suppresses the user confirmation prompt. This
option is useful for scripting operations.
Parameters:
PacketBufferOverallocationRatio : (Optional) This sets the packet buffer
over-allocation ratio for all VC-Enet modules in the VC domain. Setting
the parameter adjusts the capacity of the hardware to absorb traffic
bursts without data loss and thereby potentially increases data
throughput. Valid ratios include 1, 2, 3, & 4, where 1 is the default.
NOTE: PacketBufferOverallocationRatio is applicable only for
HP 1/10Gb VC-Enet and HP 1/10Gb-F VC-Enet
modules and deprecated for all other module types.
FlowControl : (Optional) "auto" means VC Ethernet modules will
decide automatically based on port type if flow control should be
enabled or disabled. "on" means flow control is enabled for both
RX/TX directions for all ports. "off" means flow control is disabled
for both RX/TX directions for all ports.
NOTE: Port-level Flow Control does not apply to ports on which FCoE is
configured. FCoE utilizes priority-level FlowControl only when
configured.
Examples:
- Modify the packet buffer overallocation ratio to 2:
->set advanced-networking PacketBufferOverallocationRatio=2
- Turn off flow control on all ports in the VC domain:
->set advanced-networking FlowControl=off
And last input from Mark:
From what I remember the CLI means this.
Auto is how VC is by default. Only the downlinks advertise Tx/Rx flow control. Uplinks/Stacking Links off
On is no change for the downlinks --- they still advertise Tx/Rx. The uplinks depending on the speed --- may be “ON” Tx/Rx (10Gb) or “advertise” Tx/Rx (1Gb)
Off is no Tx/Rx anywhere
************
Comments?
↧
Challenge in port mirroring configuration on Cisco switch connected to VC
Ramu had a customer question:
************
Hi Champs
Need your help,.
We are working on a large customer requirement where production network traffic needs to be mirrored to another chassis through Cisco external switch. Customer is not interested in doing port mirroring on VC level as he needs to send data to VM in separate Chassis.
***********
Reply from Vincent:
***********
You already have the proper solution: put the analyzer on a rack server, not on a blade connected to VC modules.
VC will never forward Ethernet unicast frames to a blade server when it knows the destination MAC address in the frame resides elsewhere than the blade server in question. All you could get captured on the blade server is broadcast frames and unicast frames where the destination MAC is unknown to VC, or is the blade server itself.
You just CANNOT put a network analyzer on a blade server connected to VC to capture traffic coming from the outside, there is no configuration that can make it work.
*************
Comments?
↧
Interconnect module removal from VCManager
Question from Ajith on Virtual Connect:
***************
We can we delete physically removed interconnect module from Virtual Connect domain?
***********
Info from Fred:
***********
Hi Ajith,
The VCM UG http://h20564.www2.hp.com/hpsc/doc/public/display?docId=c04383983 has a section on removing or replacing VC modules on page 275.
************
Comments?
↧
↧
Virtual Connect Manager Freezes
Please take a look at the attached screenshot.
System is a C7000 g2 with two HP VC Flex-10 Enet Modules installed.
It was working fine until two days ago. Now I get as far as the screenshot and it freezes. The menu won't load and I can't progress. Have tried various browsers and computers. Same problem.
I have logged into the HP BladeSystem interface which still works and I have reset the two modules but that has made no difference.
VCM is version 3.51. Flash is up to date.
Any ideas? Thanks.
↧
Re: Virtual Connect Manager Freezes
We've just eperienced the same issue, it appears that the SWF files no longer want to run after the 6th of May for some reason. Some sort of hitherto unknown time bomb?
Anyway, this can be worked around by replacing the files in your local flash cache; download the following from adobe:
http://fpdownload.adobe.com/pub/swz/flex/4.1.0.16076/spark_4.1.0.16076.swz
http://fpdownload.adobe.com/pub/swz/flex/4.1.0.16076/rpc_4.1.0.16076.swz
http://fpdownload.adobe.com/pub/swz/flex/4.1.0.16076/osmf_flex.4.0.0.13495.swz
http://fpdownload.adobe.com/pub/swz/tlf/1.1.0.604/textLayout_1.1.0.604.swz
http://fpdownload.adobe.com/pub/swz/flex/4.1.0.16076/sparkskins_4.1.0.16076.swz
http://fpdownload.adobe.com/pub/swz/flex/4.1.0.16076/framework_4.1.0.16076.swz
http://fpdownload.adobe.com/pub/swz/flex/4.1.0.16076/framework_4.1.0.21490.swz
Rename the files like so:
framework_4.1.0.16076.swz -> 871F12AF0853C06E4EB80A1CCAB295CEADBB817A.swz
framework_4.1.0.21490.swz -> F78F74378B1552FF9A1725155D1B43BA54BE9094.swz
osmf_flex.4.0.0.13495.swz -> C3306B26751D6A80EB1FCB651912469AE18819AB.swz
rpc_4.1.0.16076.swz -> 6DDB94AE3365798230849FA0F931AC132FE417D1.swz
spark_4.1.0.16076.swz -> 6344DCC80A9A6A3676DCEA0C92C8C45EFD2F3220.swz
sparkskins_4.1.0.16076.swz -> 440AE73B017A477382DEFF7C0DBE4896FED21079.swz
textLayout_1.1.0.604.swz -> 381814F6F5270FFBB27E244D6138BC023AF911D5.swz
Then copy them into your local flash cache that, on windows, can be found under %appdata%\Adobe\Flash Player\AssetCache\<randomly named dir>, overwriting the files there of the same name. After restarting your browser the SWF modules should load correctly.
Alternatively you can upgrade to VC 4.x but we weren't in a position to do that for all our enclosures.
↧
Re: Virtual Connect Manager Freezes
Excellent! That worked perfectly. No idea how you worked out that solution!
I will updated my VCM when I get the chance as I expect it will be an ongoing problem.
Thanks again.
↧
Virtual Connect (VC) legacy to expanded mode question.
Yaniv had a customer question:
*************
Hi experts
I have Customer that has 5 c7000 enclosures with Virtual Connect FlexFabric that is on VM4.31 version
The Customer has a problem that he can’t add any vlans to nics in the profile. The Custoemr has 28 Vlans on the NIC.
I find that the parameter of VLAN CAPACITY is on legacy VLAN capacity and that why the CU have the problem.
I want to change it to Expanded VLAN capacity.
Can I do it without any reset to the VC, Domain? It is a production environment.
************
Reply from Fred:
***********
Virtual Connect Flex-10 and more recent VC modules support expanded capacity.
Changing from legacy to expanded is non-disruptive.
***********
Comments?
↧
↧
Cisco MDS 9148 supported with VC 8gb 24-ports FC module
A question from Peter:
************
HI Experts,
Anybody knows is the cisco Model SAN Type Cisco MDS 9148 is supported with our HP VC 8Gb 24-Port FC Module or not ?
i checked on http://h10032.www1.hp.com/ctg/Manual/c02617859 and the document says N/A
So I guess it was not tested at publishing of that document ?
Now we got a major client asking if it is supported or not?
*************
Reply from Alex:
**********
It is supported. Our internal support matrix has it listed.
************
Comments?
↧
HP Blade C7000 - Stacking Redundancy Error Bay5 and Bay6
Hi all,
I'm having the following:
Blade C7000
HP VC 4.31
Bay1. VC Flex-10 Enet Module
Bay2. VC Flex-10 Enet Module
Bay5.VC FlexFabric 10Gb/24-Port Module
Bay6.VC FlexFabric 10Gb/24-Port Module
Stacking Links has a Warning about Redundancy Status.
Current Link state.
10 Gb enc0:Bay1:PortX7 enc0:Bay2:PortX7
10 Gb enc0:Bay1:PortX8 enc0:Bay5:PortX8
10 Gb enc0:Bay2:PortX8 enc0:Bay6:PortX8
As shown, there's no link between Bay5 and Bay6. Based on documentation, there should already be an Internal link between bay5 and bay6. Why doesn't they show up ?
THere's note below the connection.
NOTE:
Port X7 and X8 may connect to the internal link between horizontally-adjacent HP VC Flex-10 Enet and HP VC FlexFabric 10Gb/24-Port modules.
Ports X11, X12, X13, and X14 connect to the internal link between horizontally-adjacent HP VC Flex-10/10D Modules.
Ports X9 and X10 connect to the internal link between horizontally-adjacent HP VC FlexFabric-20/40 F8 Module.
Logs shows:
Domain state NO_COMM : Stacking Links are not fully connected, Previous: Domain state NO_REDUNDANCY, Cause: Stacking Connectivity Error: enc0:iobay6 does not have a stacking path to enc0:iobay5
Domain state NO_REDUNDANCY : Stacking Links are not redundant, Previous: Domain state NO_COMM, Cause: Stacking Redundancy Error: if enc0:iobay1 is removed, then enc0:iobay6 does not have a stacking path to enc0:iobay5
Question:
1. Where else can I find cause of missing link between bay5 and 6 ?
2. Can I manually run cable on port X7 between 5 and 6 ? I don't feel this would be necessary, since I may add couple more module that will need to link bay5 and 6 out.
Thank you!
Giang
↧
16G FC physical function in 20G CNA ?
Bibhu had a customer question on Virtual Connect physical function capabilties:
***************
Hi,
A customer of mine is asking whether they can carve a 16G FC path and have the other 3 Ethernet physical functions with a max of 4G bandwidth in the 20G CNA, using VC FF 20/40 Module.
My understanding is that the max supported is 8G. Pls advise.
***********
Reply from Vincent:
***********
Right, FCoE connections are limited to 8Gb when you're using an FC fabric on the uplink side of the VC FlexFabric -20/40 F8 module. The module FC uplinks are 8Gb anyway, and FCoE traffic on a downlink will not span multiple FC uplinks, so it wouldn't do any good to be able to carve more than 8Gb for FCoE on the downlinks.
If you're using multi-hop FCoE, i.e. your FCoE connection in the server profile is using an FCoE network on the module uplinks and not an FC Fabric, then you can carve up to 20Gb for the FCoE connection.
***********
Comments?
↧