Dec 13th, 2011 Expert Day | HP-UX & Integrity Virtualization

[Jack007]

Q1:Does vpar6.0 (VSP) need a host OS (Like Integiry VM）？



Q2:Can I assign dedicated phsical NIC to a vpar? If I have 16 NICs of a BL890c i2 and 3 vpar on it, can I give every vpar 3 NICs?



Q3:In Vpar6.0 Admin guide, it says "Currently, the NPIV feature is available only through Qlogic 8GB HBA". but my customer's FC HBA card is Emulex LPe1205 and I have checked in its quickspecs, this card support NPIV feature. So why not Vpar 6.0 NPIV feature support Emulex mezzanine card?



Q4:where can I find more docs about Vpar6.0

Thanks ! 

[Doug_Lamoureux]

---

@Jack007 wrote:

Q1:Does vpar6.0 (VSP) need a host OS (Like Integiry VM）？

Yes -  and the OS must be running the Septermber 2011 update to HP-UX 11i v3


Q2:Can I assign dedicated phsical NIC to a vpar? If I have 16 NICs of a BL890c i2 and 3 vpar on it, can I give every vpar 3 NICs?

Not with vPars 6.0.


Q3:In Vpar6.0 Admin guide, it says "Currently, the NPIV feature is available only through Qlogic 8GB HBA". but my customer's FC HBA card is Emulex LPe1205 and I have checked in its quickspecs, this card support NPIV feature. So why not Vpar 6.0 NPIV feature support Emulex mezzanine card?

We do not have the necessary driver support at this time.

Q4:where can I find more docs about Vpar6.0

 You can find the manual and release notes on the BSC site:

http://h20000.www2.hp.com/bizsupport/TechSupport/DocumentIndex.jsp?lang=en&cc=us&taskId=101&prodClassId=10008&contentType=SupportManual&docIndexId=64255&prodTypeId=18964&prodSeriesId=4146102

Thanks ! 

---

[Alan_153]

Q1:Does vpar6.0 (VSP) need a host OS (Like Integiry VM）？

The VSP is like the HPVM Host.  The VSP itself doesn't require a HOST, but "acts" like the host for vparsv6.  It sounds like you found it, but check out the release notes at:

http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c03091131/c03091131.pdf

Section 5, Page 13 talks about the VSP requirements for the OS being 11iv3 Sept 2011 (1109) along with some patches.

Alan Hymes

[Jack007]

Dear Doug_Lamoureux,

Appreciated for your answer!

About my question 2, my customer need 3 NICs each vpar to deploy oracle 11g RAC, how should I configure the vparnet?

[Craig_Fuget]

---

@Jack007 wrote:

Q1:Does vpar6.0 (VSP) need a host OS (Like Integiry VM）？



Q2:Can I assign dedicated phsical NIC to a vpar? If I have 16 NICs of a BL890c i2 and 3 vpar on it, can I give every vpar 3 NICs?



Q3:In Vpar6.0 Admin guide, it says "Currently, the NPIV feature is available only through Qlogic 8GB HBA". but my customer's FC HBA card is Emulex LPe1205 and I have checked in its quickspecs, this card support NPIV feature. So why not Vpar 6.0 NPIV feature support Emulex mezzanine card?



Q4:where can I find more docs about Vpar6.0

Thanks ! 

---

Q1: Yes, it does.

Q2: Not directly in this release; however you can manually ensure that each NIC port is used by one and only one vPar, and that the port is not used by the host OS.

Q3: Thanks for the input. 

Q4: More documents are coming

[Jack007]

Alan,

Thanks a lot! 

[Jack007]

Does VPAR6.0 support APA well ? 

Does VPAR6.0 supports other  adapters such as scsi and lan, or it only supports avio_stor and avio_lan?

[Craig_Fuget]

---

@Craig_Fuget wrote:

---

@Jack007 wrote:

Q1:Does vpar6.0 (VSP) need a host OS (Like Integiry VM）？



Q2:Can I assign dedicated phsical NIC to a vpar? If I have 16 NICs of a BL890c i2 and 3 vpar on it, can I give every vpar 3 NICs?



Q3:In Vpar6.0 Admin guide, it says "Currently, the NPIV feature is available only through Qlogic 8GB HBA". but my customer's FC HBA card is Emulex LPe1205 and I have checked in its quickspecs, this card support NPIV feature. So why not Vpar 6.0 NPIV feature support Emulex mezzanine card?



Q4:where can I find more docs about Vpar6.0

Thanks ! 

---

Q1: Yes, it does.

Q2: Not directly in this release; however you can manually ensure that each NIC port is used by one and only one vPar, and that the port is not used by the host OS.

Q3: Thanks for the input. 

Q4: More documents are coming

---

Just a bit more on Q2 - my workaround  is to create a number of vswitches, one per port and to make sure _not_ to share them.  I also make sure that the ports I use for the vPars are not in the VSP's netconf file (as lan0 or something).  Then, I just manually ensure that each vPar gets all of each vswitch that they are connected to.  This can be verified easily if you install the new VSMgr GUI.  The networks tab will show 1 line each vswitch and on to each network port.

[Craig_Fuget]

---

@Jack007 wrote:
Dear Doug_Lamoureux,

Appreciated for your answer!

About my question 2, my customer need 3 NICs each vpar to deploy oracle 11g RAC, how should I configure the vparnet?

---

I would create a vswitch for each NIC, and then add 3 networks/switches to each vPar. 

[Doug_Lamoureux]

---

@Jack007 wrote:
Dear Doug_Lamoureux,

Appreciated for your answer!

About my question 2, my customer need 3 NICs each vpar to deploy oracle 11g RAC, how should I configure the vparnet?

---

Networking in vPars v6 is configured the same way it's done on Integrity VM - 1st you create a vswitch on the VSP that is backed by a physical NIC - That vswitch is then presented to the vPar(s) - the vPar see's it as a physical NIC (lan1, lan2, etc...)  - If you want to share the physical NICs across multiple vPars you would configure each vpar with the same vswitch(s).    If you only want the physical NIC to be used by a single vPar do not configure multiple vpars with the same vswitch.  For example:

# vparnet -c -S vs0 -n 0

# vparnet -c -S vs1 -n 1

# vparnet -c -S vs2 -n 2

# vparnet -c -S vs3 -n 3

# vparnet -c -S vs4 -n 4

# vparnet -c -S vs5  -n 5

# vparmodify -p vpar1 -a network:avio_lan::vswitch:vs0  a network:avio_lan::vswitch:vs1 -a network:avio_lan::vswitch:vs2

# vparmodify -p vpar2 -a network:avio_lan::vswitch:vs3  a network:avio_lan::vswitch:vs4 -a network:avio_lan::vswitch:vs5

Please be aware that this is NOT direct access to the physical LAN interface – you are still using a virtual NIC in the vPar.

Cheers,

Doug