Vsan UCS - Nexus - questions

Similar Questions

• UCS/Nexus and FCoE Question

  Hi all:

  Can we connect a FCoE target directly to an interconnection of fabric or a Nexus 5xxx CWF in NPV mode and not connected to an FC switch for power and tissue services continue to have the initiator see the target and be able to supply storage as a result?

  Thus, the scenario would be the following:

  a server rack with an ANC connected to a Nexus 5xxx CWF in NPV mode

  OR

  a server blade a NON related to interconnection of UCS fabric (which is also a CWF) in VAN mode

  AND

  In the above cases, there is NO CONNECTION between the RTC and an FC switch.

  I will be able to have the ANC (FCoE initiator) see table of Strorage (FCoE target)?

  Thank you

  Hello

  Yes you are right. For the direct-attached storage work, you need the FI in FC switch mode. Your explanation is good and you can not storage direct-attached mode/npv end FC host mode. With FI in npv mode, it can just transfer CF applications to an NPIV switch and cannot process any request CF as FLOGI etc...

  In addition, in the UCS versions under 2.1 you need a SAN switch connected to a FI running in mode switching and storage directly connected to it, which can push the FI database zoning. Thanks to 2.1, you can have storage direct-attached without the need for any SAN switch upstream as we have zoning capabilities in UCS now.

  In short, for all direct-attached storage, you must have FI FC Switching mode.

  -Ganesh

• UCS General questions

  Hello

  I have a few questions about UCS. Do not hesitate to reply to all the of them, thank you!

  1. What are the main benefits of using FCoE vs FC to UCS?
  2. can someone point me to performance data of FCoE vs FC in the UCS?
  3. What is the difference between the fixed modules and expansion on made them other than being able to replace the plug-ins? A difference that I noticed is that when you select "Configure the Ports unified" on the FI, it says that all FI will have to reboot if you use fixed ports, but the port expansion module can restart independently. There are other differences between the fixed modules and expansion?
  4. What is the best practice for creating organizations? The limit is usually the data center? What are the benefits for the combination of several data centers (which could be in the same geographic area) into a single entity?
  5. most people configure UCS of 'root', or they always configuring everything in specific OUs they create?
  6. This setup of most of the organizations mentioned here video separates VSAN for each fabric (A & B)
  a. How are the fiber channel pass setup to accommodate the VSAN on made them?
  7. is there an advantage to having separate VSANS / VLAN ID between fabric A and B? A few examples I see use the same ID of the VSAN on both fabrics, while others are not
  8. in this diagram (link), there are 4 connections from the FIs to the San. Each FI is wired to each storage processor. In this configuration, how it the SAN be configured - ALUA?
  9 (for ESXi hosts) what are the advantages/disadvantages of creating a model vNIC for each tissue, compared to one model for each vmnic vNIC?
  10. under what circumstances should you not check "Activate the failover" for models vNIC for esxi hosts?

  11. What is the best practice for the connection of FEX-to - IOM to the title of the material > policies > policies > chassis/FEX discovered political? Many companies set it up as a Port channel or not?

  (1) complex is the best I can say.  For us, we decided to stay with native CF that we already have the buildout FCC native. Our network group is for us to move to FCoE (we use N5K for our data center fabric switches).  But for our SAN group, we all all comfortable with native CF right now and didn't want to convolute our UCS deployment at a crossing FCoE.

  2) UCS FCoE, comes from the back of the interconnection.  There is no difference in performance. Once FC lands to interconnection, interconnection opens the package and converts it to the headers of FCoE, laying on a VLAN you configure (must be a VLAN unique, it will leave you not overlap with production VLAN defined in UCS already) and sends it to the VIC. The VIC then opens the package FCoE and hands to the BONE as if it were native fiber channel.  The difference in performance between the two is negligible.

  (3) the fixed module is controlled a.  When you change the unified ports any interconnection charges.  When you load a plug-in only the plug-in restarts if you change its setting unified ports.  Another thing to note when buying a plug-in you get you license additional ports that can be used everywhere where (it doesn't have a plug-in port license).  Other than that, the performance and the functionality is the same as fixed.

  (4) create us organizations for each type of OS because different groups managing different OS'. For example, we have an org for VMare, one for Linux, for windows, and then a special one oracle for our DBAs. We then assign permissions via AD to local so the groups have access to their servers with a minimum of effort on our part.

  (5) while root except if necessary below (haven't run this yet)

  (6) as a "best practice", you should always have two separate fabrics. UCS must put the FCoE VLAN fabrics.  To avoid confusion, it is best to put each fabric in its own VSAN. For example, we run our SAN fabric production in VSAN around 10 and FCoE VLAN 3210.  SAN fabric B is to VSAN 11, FCoE VLAN 3211.
  (6A) the prospect of a MDS/Nexus that you put in place a "f-port trunk." Newspapers of the UCS in the fabric as a nport with NPIV (ucs connects and then tells the fiber switch "also, these additional WWPN are also on this port, I just connected to, send their data to me").  You put the f-port channel you create on the mds/nexus device in the VSAN you want (in my example, 10 for A VSAN and VSAN 11b). Also to note: you cross never interconnection fabrics up to your SAN infrastructure. InterConnect will have a port channel to A SAN fabric and
  Interconnect B to B SAN fabric.  Consider the interconnections as HBA on a normal server.  You do not mix the tissues A and B.  That's how you get redundancy SAN.

  (7) see above.  I strongly recommend that create you separate VSANS for each fabric.  This will ensure that nothing ever gets through.

  (8) If you connect directly FIs them processors SAN you must activate the switch for on the UCS fabric mode.  You won't be able to connect the UCS for all fiber switches once do you it, and all the zoning must be done in UCSM. It does not work (we do in our test environment), be aware that it fits well beyond an area.  You would provide the same as before your San.  Activate ALUA.  The first port on each SP would go to FI - A and the second port on each SP would go to the FI - B.  Then the appropriate area in UCSM.

  (9) our ESXi hosts get a minimum of six models of vNIC.  We create an A, B mgmt, A vmotion, vmotion mgmt B, features and data B.  As you can guess, pin us each model to a cloth and do not switch.  This gives us a ton of flexibility within ESXi.  Furthermore, I believe that it is the 'best practice' is not to use the failover fabric for VMware. (If you chose, you will get alarms in vmware indicating that there is no redundancy nic.  There is an advanced disable setting.  Do not have casual, you would need to google that.). I recommend you to make a template for each tissue and keeping them up-to-date as a pair.

  (10) I wouldn't use fabric failover for everything related to VMware, VMware does a great job of redundancy management.  Let it do its work.  We only use the fabric failover for windows and Linux host running the SCU.

  (11) Yes set up as a port channel.  Before you connect the chassis as you can not no interruption of service do this later.  When a frame is automatically recognized, port channels or groups of pine are configured behind the scenes.  You cannot change these directly. Therefore, if you switch to port channels later, the only way to 'rebuild' is to re - emphasize the chassis, which will be the frame and everything in it in offline mode.  Bad day.  If you do not port channels, use the vETH blade interfaces get pinned to an uplink (following the defined process of cisco, google for UCS once again pinning get more information).  The channels of port pine servers to the interface port-channel instead of the uplink interface which makes it much more stable and resistant to link failure.  Pinning, if memory is, if a link to a server is pinned breaks down, the interface of the server is down.  It is not "repin" the link.  Thus, for example, if you do not use channel ports and a server gets pinned to uplink 1/1, and it breaks down, the server will see this interface out of service.  With a port channel, as long as a link in the chain is in place, the server will be displayed.

  I hope this helps. I typed this from my ipad - do not hesitate to respond and I'll try to clarify.

• ID OF THE VSAN UCS

  Hello

  I have a small question on the VSAN ID and hope you guys have an answer for that.

  When you create a VSAN in UCS FI, ability I use the same VSAN ID created in the MDS switch upstream or can I use a different ID?

  See you soon,.

  Kisanka

  Yes, the VSAn Id should be the same as the MDS upstream.

  . / Afonso

• CIsco UCS FI question license (2,23e)

  Hello

  I have system UCS running firmware version 2.2 (3rd) and after the application of the licenses of four additional ports for each fabric of interconnection, I got a warning that tissue-period of grace entered interconnection.

  We have these warning although we are not on the supply of all ports (e.g. absolute amount = quantity). Please check the attached screenshot.

  I want to check if anyone had this problem before and if it of the bug or not.

  Thank you

  Mohammad

  https://Tools.Cisco.com/bugsearch/bug/CSCui19338/?reffering_site=dumpcr

  Perhaps?

• Integration of UCS with vSphere - invalid URI error

  Hello

  I fight to integrate the Manager UCS with vSphere vCenter.

  The extension of vCenter is exported from the UCSM without error, but when you try to install the plugin via vCenter plug-in Manager I get this error:

  "Invalid URI: the URI is empty.

  I found this on the web, but this brings me any more to solve the problem:

  http://terenceluk.blogspot.com/2010/11/error-message-when-clicking-on-download.html

  Any tips?

  Kind regards
  Radek

  Hi Radek,

  So as long as the file is saved in the folder plugins in vCenter, and they appear in the list of plugins, then that's all you need to do. You don't need to click on the 'Download and install' button as this will not work, because you have met. This is because it is a document XML allows to activate and validate your vCenter server and communication between UCS/Nexus 1000v. There is no practical application which is executed and must be installed.

  Hope that helps to clarify

  Thank you

  Michael

• Balancing of VMware with Nexus 1000v

  With the vmware puts vDS or vSS, I see many designs use the asset-liability approach for binding rising consolidation of NETWORK cards, IE a vmnic is active will fabric and a vmnic is passive will fabric B. This setting is configured in vSphere.

  Se this article: http://bradhedlund.com/2010/09/15/vmware-10ge-qos-designs-cisco-ucs-nexus/

  Is this correct, that we can put in place a regime with the 1000V? All the network is on the 1000V config, and as far as I know, we can only configure the uplink in these 3 modes:

  1. LACP 2. vPC-Host Mode 3. vPC-Host Mode Mac pinning

  and they are all 'active' based.

  Post edited by: Atle Dale

  Yes.  All uplinks are used.  Each VM virtual interface is pinned to one of of the uplinks.  If one uplink goes down, all interfaces pinned gets dynamically likes to remaining uplinks.  A mac address will only see on a single interface at a time.  This is how MAC pinning prevents STP loops.

  Robert

• Change of address IP UCS Central

  Hello

  Anyone know if it is supported?

  Thank you very much

  Hi Frank

  It's hard, why?

  -Register a domain UCS, does not support DHCP for obvious reasons.

  -You can not change or swap the IP addresses used by Cisco UCS Manager in a domain that is registered with Cisco UCS Central. If you need to change or swap this IP address, you must first cancel the Cisco UCS Central domain registration. You can re-register the domain of the Cisco UCS after changed or exchanged the IP address.

  -If you use a hostname instead of IP address, you must configure a DNS server. If the Cisco's UCS domain is not registered with Cisco UCS Central or DNS management is defined at the local level, configure a DNS server in Cisco's UCS Manager. If the area of Cisco UCS is registered with Cisco UCS Central and DNS management has the overall value, configure a DNS server in Cisco UCS Central.

  So maybe if you used a DNS hostname, instead of an IP address to register a UCS domain, you can change the IP address in the DNS (and IP address in the center of the UCS).  Question is: How can you use UCS Manager to do a reverse DNS lookup and choose the IP address has changed. Write off / registry is not the solution.

  Summary: I don't think that there is a decent solution.

  Walter.

• Please tell me how to connect with FCoE mini-UCS and Nexus N5k (N5K-C5548UP-B-S32)

  I have UCS 5108 chassis of Server Blade (with two blades B200 M3 inside and two FI 6324) connected to the switch of Nexus N5k 4x10GE links.

  I want to set both aggregated channels FCoE (vFC) of chassi to pass.

  All the instructions I've seen mention that UCS-mini needs to be mode end-host FC, but version 3.0 UCSM does not support this mode, what should you do?

  I put in place following the www.cisco.com/c/en/us/support/docs/switches/nexus-5000-series-switches/116248-configure-fcoe-00.html instructions

  But after the installation, I get an error on the fabric: FCoE or FC uplink is down on Vsan 500;

  And the Nexus program: Vsan 500 is declining (pending flogi)

  When I ' display interface vfc 1 ' CLI nexus I don't see "Trunk VSAN (high)", but

  "VSANS trunk (boot) (500).

  Unfortunately, I can't find the technical notes for mini UCS (with UCSM v. 3.0), so maybe you can promt guides or suggestion for this?

  Thank you in advance!

  Same question here.

  TIA

• Nexus 1000v, UCS, and Microsoft NETWORK load balancing

  Hi all

  I have a client that implements a new Exchange 2010 environment. They have an obligation to configure load balancing for Client Access servers. The environment consists of VMware vShpere running on top of Cisco UCS blades with the Nexus 1000v dvSwitch.

  Everything I've read so far indicates that I must do the following:

  1 configure MS in Multicast mode load balancing (by selecting the IGMP protocol option).

  2. create a static ARP entry for the address of virtual cluster on the router for the subnet of the server.

  3. (maybe) configure a static MAC table entry on the router for the subnet of the server.

  3. (maybe) to disable the IGMP snooping on the VLAN appropriate in the Nexus 1000v.

  My questions are:

  1. any person running successfully a similar configuration?

  2 are there missing steps in the list above, or I shouldn't do?

  3. If I am disabling the snooping IGMP on the Nexus 1000v should I also disable it on the fabric of UCS interconnections and router?

  Thanks a lot for your time,.

  Aaron

  Aaron,

  The steps above you are correct, you need steps 1-4 to operate correctly.  Normally people will create a VLAN separate to their interfaces NLB/subnet, to prevent floods mcast uncessisary frameworks within the network.

  To answer your questions

  (1) I saw multiple clients run this configuration

  (2) the steps you are correct

  (3) you can't toggle the on UCS IGMP snooping.  It is enabled by default and not a configurable option.  There is no need to change anything within the UCS regarding MS NLB with the above procedure.  FYI - the ability to disable/enable the snooping IGMP on UCS is scheduled for a next version 2.1.

  
  
  This is the correct method untill the time we have the option of configuring static multicast mac entries on
  
  the Nexus 1000v.  If this is a feature you'd like, please open a TAC case and request for bug CSCtb93725 to be linked to your SR.
  
  This will give more "push" to our develpment team to prioritize this request.
  Hopefully some other customers can share their experience.
  Regards,

  Robert

• Trouble seeing vhba of UCS on Nexus 5 k

  I have UCS > 6100 (a/b) > nexus 5 k (a / b) > emc vnx 5300.  I have two installed VSAN and my switch can see two ports on my vnx5300 with show flogi. However, my 5ks don't see the vhba on the ucs system. I have two vhba configured, a tissue in one and the other for fabric b. From now the nexus is connected to the 6100 as a port channel. Ports-channel is configured as a trunk. Now, this makes me think that this is where the vsan communication is a failure. The connection to the 6100 and nexus needs to be something more than a simple trunk?

  Dennis

  Dennis,

  I hope that you give on the Fiber Channel uplinks to your N5K and not Eth/FCoE (there no FCOE Multihop yet).

  Assuming that you have your 6100's with CF logs to your N5Ks and the N5K "npiv' enabled, you should see the flogi to your host vHBAs.

  The CLI of 6100 you can also 'connect nxos' and 'show flogi npv' which will show you the CF sessions between your 6100 and N5K - I find where you problem.

  All of your SAN connections would travel on the module of Fibre Channel 6100, up to Channel N5K Fiber module, then to your table.  You can always trunk and channel ports SAN your rising FC.  Just to confirm that this is what you did and not using uplink Ethernet to your N5K waiting to work.

  Can you provide a diagram or a detail the exact topology (Eth, FC etc)

  Kind regards

  Robert

• Configuration of Nexus 5 k for SAN switching between UCS 6100 s and NetApp storage

  We strive to get our configuration of the UCS environment make start SAN LUN on our NetApp storage. We have a pair of Nexus 5Ks with the company / SAN license on them and the 6-port FC module. What to do to configure Nexus 5000 to operate as the SAN switch between the target storage and environment of the UCS?

  You see the 6120 WWPN on the N5Ks? What is the result of 'show flogi database' and 'show int br '.

  You have mapped the FC ports on the 6120 for the proper VSAN?

  NPIV is enabled?

• UCS environment vSphere 5.1 upgrade to vSphere 6 with Nexus 1000v

  Hi, I've faced trying to get the help of TAC and internet research on the way to upgrade to our UCS environment, and as a last thought resort, I would try a post on the forum.

  My environment is an environment of UCS chassis with double tracking, tissue of interconnections, years 1110 with pair HA of 1000v 5.1 running vsphere.  We have updated all our equipment (blades, series C and UCS Manager) for the supported versions of the firmware by CISCO for vSphere 6.

  We have even upgraded our Nexus 1000v 5.2 (1) SV3(1.5a) which is a support for vSphere version 6.

  To save us some treatment and cost of issuing permits and on our previous vcenter server performance was lacking and unreliable, I looked at the virtual migration on the vCenter 6 appliance.  There is nowhere where I can find information which advises on how to begin the process of upgrading VMWare on NGC when the 1000v is incorporated.  I would have thought that it is a virtual machine if you have all of your improved versions to support versions for 6 veil smooth.

  A response I got from TAC was that we had to move all of our VM on a standard switch to improve the vCenter, but given we are already on a supported the 1000v for vSphere version 6 that he left me confused, nor that I would get the opportunity to rework our environment being a hospital with this kind of downtime and outage windows for more than 200 machines.

  Can provide some tips, or anyone else has tried a similar upgrade path?

  Greetings.

  It seems that you have already upgraded your components N1k (are VEM upgraded to match?).

  Are your questions more info on how you upgrade/migration to a vcenter server to another?

  If you import your vcenter database to your new vcenter, it shouldn't have a lot of waste of time, as the VSM/VEM will always see the vcenter and N1k dVS.  If you change the vcenter server name/IP, but import the old vcenter DB, there are a few steps, you will need to ensure that the connection of VSM SVS corresponds to the new IP address in vcenter.

  If you try to create a new additional vcenter in parallel, then you will have problems of downtime as the name of port-profiles/network programming the guestVMs currently have will lose their 'support' info if you attempt to migrate, because you to NIC dVS standard or generic before the hosts for the new vcenter.

  If you are already on vcenter 6, I believe you can vmotion from one host to another and more profile vswitch/dVS/port used.

  Really need more detail on how to migrate from a vcenter for the VCA 6.0.

  Thank you

  Kirk...

• Question of design around VSAN

  We left the Nexus 5 K for our CF to switching to start our UCS blades. I wonder what the best design is for the following situation.

  Active / Active NetApp HA pair.

  FC 2 ports target of each head of NetApp in the HA pair to the SAN (Nexus 5 K in our case) switches.

  1 port of each head of NetApp to each of the 5010 s nexus.

  Should I 1 VSAN on each 5010 with each head of NetApp in there or should I have 1 common VSAN between the 5010 s?

  Any comments on this would be appreciated.

  A recommended good is to use different numbering on each fabric VSAN.

  Fabric (or 1) odd numbered VSAN use (3, 5, etc.)

  Fabric B (or 2) use even mumbered VSAN (2, 4, etc.)

  It provides a visual cue for the operator to let them know what fabric, they make changes. In addition, in the case of inadvertence ISL to connect between the two fabrics that you merge your A and B fabric zonesets (I saw it happen). Do not "put any active trading in VSAN 1 for the same reason.

• Question of VLAN native of UCS

  All,

  I have a problem that I can not just wrap my mind autour.  We have UCS setup in a lab with 2 interconnections connected to 2 nexus switches 5510.  The nexus switches are passed to the network via a Switch 4900 m.  All circuits are configured and tested as functional. All routing is configured and confirmed.  I have a problem in UCS, which is confusing to me.  In the lab, I kept the VLAN native to the vlan1.  I have the Setup VLAN 2-10 on all switches test and interconnections.  I created a service profile that contains 1 network card and placed it in the VLAN 7.  I installed Windows 2008 on a blade using this service profile.  In the operating system I statically IP'ed the NIC for the schema used in VLAN 7.  The OS, I cannot ping another device located in the vlan 7.  Also, I can't ping a host on a different VLAN.  If I place a check on the VLAN 1 as the vlan native I still cannot ping anything.  If I place the audit for vlan native to vlan 7 I can ping hosts in the same vlan, as well as outside of the vlan.  So why should I place vlan 7 as the vlan native when all my boxes are set up in the vlan 1 is the vlan native?

  Thanks for any help,

  Ken

  Ken,

  When you allow some VLANs on your Service vNIC profile you will need to set the VLAN native. This is because the way you have configured currently you are only "allowing VLAN 15', but you're not marking it.   It will work fine for ESX or Linux which allows to assign the dot1q tag to the host.  With Windows unless you have specific drivers doing the marking for you, you will need to do it at the level of the vNIC in UCS.

  Two ways to see this in action.  When you create a service profile in the 'Basic' - not 'Expert' method, you will need to choose a single VLAN for your interfaces.  This will treat interfaces about like an "access Port".  Conversely, when you use the "Expert mode you select the vNIC as a trunk, in which you" will allow to "all VLAN you acceding them as to, like this is the method you did.»

  For a Windows operating system, set the VLAN natively for the VLAN you want to access and you'll be gentle.  Unchecking this option button that "VLAN native" is allowing traffic to cross out of UCS on the VLAN native VLAN 1, your network - it is therefore MAC appears on other fabric under VLAN1

  Kind regards

  Robert