vRA 7 load balancing with embedded active / standby vRO

Similar Questions

• T60 - problems with long active standby

  Hi all

  Hopfully, can someone point me in the right direction with a problem I encounter with one of the ThinkPad T60.

  Here's the run down:

  We did 3 of the T60 (type 6369-A56) last December (online sales of web). 1 of the three displays problems when dealing with the day before. If the laptop goes into sleep mode during more than anywhere between 30 minutes to 1 hour that the laptop is not repeated mode standby when you open the lid of the laptop or use the Fn function key. When two approaches to bring the laptop leaves standby mode the Moon cresent LED stays on, the HARD Drive LED turns on for a brief second and then turns off, the power light and strong say, the screen remains blank at any time. When this happens, you can hear/feel the disk HARD turns to the top and doing something and you can also hear the CPU fan spin in the action as if it came out the day before. Does not seem to have taken power to make a difference with this problem.

  I already talked to Lenovo support RELATIVE to this subject and they believe it's a software problem. I did a few more tests to try to determine that this is what I did...

  I took a Norton Ghost image of one of the 2 working T60 HDD (Eve works perfectly on two other laptops) and used this Image for ghost in the T60 defective HARD drive, this gave me two identical laptops (hardware and software). Power for laptop mode is as follows, turn off the display @ 5 minutes, switch Mode suspend @ 15 minutes. I left computers laptops only at exactly 09:05, 5 minutes later so empty screens, @ 15 minutes later, the two went in standby mode. 45 min after sleep mode had struck I returned to two laptop computers and for laptops (at the same time), use the Fn key and the defective T60 does not return from the day before and the same symptoms as mentioned earlier it happened.

  After this I decied to see what would happen if I changed the HARD drive between laptops (from the laptop failed drive in the laptop to work), failed hard drive of the laptop working in laptop. After the passage of the HARD drive, I ran the same tests as above. Yet once, 45 minutes after the day before had been very active, I used the Fn key to get out of sleep mode and the laptop with the failed hard drive from the laptop worked directly at the exit of the batt, but the hard drive of the laptop from work to the faulty laptop displayed the same problems as mentioned above thus excluding the software as the cause of the problem.

  My question is, knowing all this, where is the problem likely to be? Is - it's going to be RAM related or associated with Board?

  The result of likley going to be back to Lenovo for repair?

  See you soon
  Chris

  Hello

  Yes, after trying one number of other things I determinined it was the motherboard at fault. Introduced a job guaranteed with Lenovo and returned to the unit for repair. Less a week we had it back, running without problem (they replaced the sys Board).

  See you soon
  Chris

• VM simple Load Balancing with 4 physical network adapters

  I have one main server with 4 cards of physical network dedicated to this server in a vSwitch. I wanted to know the proper way to balancing configuration in vSphere to allow a theoretical throughput of 4 Gbit/s. The options I have, as you know:

  1. Port according to ID - this the method by default when you have 2 or more physical NIC connected to a virtual switch. The VM traffic is realized on a physical NETWORK card based on the VMs virtual port ID and is incremented in the style of round robin. So if you have 2 NICs physical network, the first drop of will traffic VMs that the first NIC, the second VM shuts down to the second NIC, third comes out first NETWORK card and operating system on - host ESXi does not resemble the traffic so if VMs 1,3,5 are heavy network users they will go on the same NETWORK adapter even if the second NETWORK card can be totally unused
  2. Address based MAC - based similar to port, but the physical NETWORK adapter is selected according to the MAC address of virtual machines
  3. IP hash database - the physical NETWORK adapter is selected based on the starting and destination IP address - so if a virtual machine connect to several IP addresses that traffic will be disttibuted on all physical network cards - note this will require LACp be configured on the physical spend this ost ESXi connects to

  The first two sound me as they would give me what I want from the physical NIC selection is based on the virtual source, not the destination machine. So, since I have only a virtual computer, the LB would never use a physical NIC and I'd be stuck with 1 Gbps.

  The third option seems more promising, but I do not have much exposure to the management switch. LACP would be similar to the creation of a trunk on my 3Com switch 2848 reference? It seems likely, but I wanted to run it by the experts to make sure.

  Thank you!

  EDIT: I think I just found the killer. Cited in the manual for my switch:

  The switch does not support the aggregation of links
  Protocol LACP (Control), which is specified in the IEEE standard
  802.3ad.

  Does that mean I'm desperate to do my 4 Gbps with the swithc and I MUST get a LACP switch?

  You will need either LACP or EtherChannel on your switch upstream to support IP hash configuration, install, configure, manage Volume 1 book for ESXi 5.0. With hash IP, MAC address of the virtual computer may appear on any of the physical interfaces available for the vSwitch (at least for the outbound traffic), which would cause problems on the side of the physical switch unless a method of aggregation (LACP or EtherChannel) link type is enabled.

  Mike

• Configuration of AppsTier (CFP, load balancing) of nodes R12 Multi with RAC 2 nodes

  Dear all,
  
  I installed and configured applications R12 with Oracle RAC 2 nodes. (APP1, DB1 and DB2).
  
  Now, I need to add a layer of application (APP2) to configure high availability, balancing load and PCP.
  
  So, I cloned (adcfgclone.pl appsTier) application for APP2 layer (not shared, the two file systems are different). When I run the application it doesn't knot APP2 App1, more more simultaneous manager works only on a node (can be due to APPLCSF does not point to the mounting NFS share folder), although I've defined managers primary school-children also.
  
  If I run the autoconfiguration of APP1 application service will only work on APP1 and even if I do for APP2, then application works with APP2. I need to have configuration CFP, balancing the load (via DNS if possible) and the high availabllity for the Application tier.
  
  When I start the application, it will start OPMN and APPS Listener on the nodes, but simultaneous manager does not start on both, to a single node.
  
  Is this requires having both APPS name of the node is the same?
  
  Thank you very much
  
  Ravi Purbia
  
  Published by: Purbia delighted on June 9, 2009 18:04
  
  Published by: Purbia delighted on June 9, 2009 18:05
  
  Published by: Purbia delighted on June 9, 2009 18:07

  Ravi,

  If I run the autoconfiguration of APP1 application service will only work on APP1 and even if I do for APP2, then application works with APP2.

  Most likely, it's something with the profile option values, then run utility adchkcfg to get an html report that lists all files and profile options that change when you run AutoConfig on each node.

  I need to have configuration CFP, balancing the load (via DNS if possible) and the high availabllity for the Application tier.

  For load balancing, you must perform the steps according to the guidelines of this document.

  Note: 380489.1 - using Load-balance with Oracle E-Business Suite Release 12
  https://metalink2.Oracle.com/MetaLink/PLSQL/ml2_documents.showDocument?p_database_id=not&P_ID=380489.1

  For the installation of the CFP, please see:

  Note: 388577.1 - using Load-balance with Oracle E-Business Suite Release 12
  https://metalink2.Oracle.com/MetaLink/PLSQL/ml2_documents.showDocument?p_database_id=not&P_ID=388577.1

  When I start the application, it will start OPMN and APPS Listener on the nodes, but simultaneous manager does not start on both, to a single node.

  Any errors in the log file of CM?

  Is this requires having both APPS name of the node is the same?

  It is not necessary.

  Kind regards
  Hussein

• DRS load balancing

  Hi people,

  Based on a few blogs now I understand now DRS load balancing based on active memory used on ESXi individual cluster nodes. Yesterday, we migrated VM to an ESXi node which pushed the memory consumed 99% but DRS seems never disturbed. I assumed that justification as the memory usage active only 11 GB of 160 GB.

  But my query is that what happens when I migrate several virtual machines to hosts that already shows alarm for 99% of the consumption of memory? If the DRS was examined only active use so why he throws alarms based on consumption, creating panic with the customer?

  Please have helped to understand the same.

  Thank you!

  Best regards

  Hari.

  Hi Hari,

  Good point.

  DRS behavior can be changed in order to study the memory consumed as well. U2c 5.5 and above, DRS considers the memory consumed as well. How much he believes is below specified KBs.

  Reference: https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2059868

  There is a KB linked to top Ko. Please throw in that as well.

  According to me, the alarm is little generic, it works even without the DRS, it seems, the alarm is designed in a way that it is not aware of this host belongs to active DRS cluster. DRS comes with the license of the company and alarm will work with basic licenses as well.

  are you suggesting that alarm should be aware of the hosts that are part of the active DRS cluster and give only the alarm according to the metric memory configured on DRS?

• The order of failover and load balancing

  Hello

  I have the following scenario. An ESXi with 4 Gbps vmnic. The questions are:

  (1) if I have a group of ports configured for 'Route based on the original virtual Port code' in the policy of balancing load, and for the same port group I the option button 'Override switch failover command"checked, where I set up 3 of the active adapters vmnic, as well as the other vmic remaining as unused adapter, the ESXi uses the policy that I have configured (in this case 'Route based on the original port code') between the three vmnic load balancing marked as active? Or he uses them in the order that they appear in the section active cards?

  (2) Suppossed, I configured the four physical switch ports in an etherchannel group to use 'Route based on the IP hash' load balancing policy. In this situation, then I configured for a certain group of port to only used two active adapters and two others as unused? In this case, ESXi should balance the load using the method hash IP but only in two active adapters? Or it is a misconfigiuration and I should not configure my nic teaming in this way?

  (3) the official setup guide says "NOTICE on IP requires the physical switch be configured with etherchannel. For all other options, etherchannel must be disabled. ». How can I I configured my virtual network, if I have a few groups of political ports based on the hash of the IP to use load balancing and another uses 'Route based on the original port code. This is the case when I for example have two management ports using the same vSwitch with four vmnic (where they are configured as an Etherchannel in the physical switch). I would port one or several groups for virtual machines that use the IP of the hash method of balancing the load and vmkernel ports por management uses only a single adapter active with no back and as "based on the source port ID" load balancing as best practices said.

  Now, the four vmnic is the same for all traffic. The physical switch ports must be configured in an etherchannel group because certain groups of ports will use the method of IP hash, but others are not. The configuration guide I said SHOULD NOT use etherchannel if I won't use the hash IP method, but I'LL use it, but only in groups of one or more ports.

  Maybe I do not share the same vmnic from this situation.

  Finally, it's a philosophical question. What is the difference between 'The route based on the source port ID' and the 'road based on the source MAC Hash' load balancing policy? What is the purpose of the second? It is assumed that if I had two different MAC address in a virtual machine, it would be because I had two different virtual cards inside the virtual machine, which would be connected to two different port ID in the vSwitch, I can use the first strategy (based on the original port code). In other words, which would be the case where I had the traffic entering the same vSwitch but port ID with different source MAC address, so I should chose the method to distinguish the Source MAC address load balancing traffic?

  Thank you.

  Guido.

  (1) as long as you override vmnic only and don't change the policy for this group of ports, he uses the policy configured at level vSwitch and use the selected interface 3 with this policy

  (2) it should work, I don't think it's a problem for the switch receive packets on a subset of the aggregation. I do not think that Etherchannel is supported (IIRC, it is a Cisco proprietary protocol, VMware only supports LACP passive, which corresponds to the Port channel world Cisco.) Trouble me if I'm wrong!)

  (3) I think that's all right, as I have explained in 2), there is no special negotiations with the consolidation of VMware, the important thing only I know is to configure the port on the side of the switch channel if you decide to use the IP hash (that will lead to important questions)

  4) (self labeled) I think it may differ in some cases individuals, as when the operating system use the same MAC address for both NICs (aggregation in-vm) or if you advertise several MAC address for the same network card (ESX in a VM for example would make for its VM). Such cases differently affect this setting.

  That is the right question, and I'm curious to know if someone wants to develop on it!

• DSwitch load balancing

  I have DSwitch 5.5 in vsphere 6.0

  Two 5.5 esxi hosts to connect to these virtual switch of two network cards.

  NICs works in IP HASH (the real switch side cisco 3570 I did port channel). Everything works fine, but I have question.

  Can I do a load balancing for a single host examle: IP hash, on the other - on the virtual port original? If say, how? I can see only one setting for all VDswitch

  

  

  Can I do a load balancing for a single host examle: IP hash, on the other - on the virtual port original? If say, how? I can see only one setting for all VDswitch

  As you can see on your screenshots, settings are for port group and not overall overall dvSwitch. You can have different settings for each physical ESXi host though, all guests must use the same algorithm as specified by the port or load balancing groups. If you want different settings by physical ESXi hist, then you will need to use the local standard and undistributed vSwitches, but we can ask the question of whether such a potentially unpredictable and inconsistent configuration is desirable.

  On a side note, you can set up several groups of ports with different parameters on a vSwitch (d). However, mixture etherchannel/LAG/LACP IP - hash based load balancing with any other balancing as virtual port ID in the same uplink / vSwitch (d) is not supported.

  In addition, it is not sensible, your acts only physical switch as a single channel for all traffic and will not be able to differentiate as several ports on an ESXi host groups could do.

• Console Proxy doesn't work is not behind the edge load balancer

  I have a strange problem. I set up a load balancer with vShield Edge. Behind the load balancer, I have two cells vCloud. The Web interface works very well for users, but trying to connect a VMRC to view remote screen it displays 'connection' and then the session is disconnected.

  Then I stop the first cell and users can establish a VMRC connection again. When I activate the first cell once again, the web interface continues to work, but not the proxy of the console. Then I stop the second cell and now elements of the console working again.

  Any tips?

  Gabrié

  Have a similar setup and had the same problem.  Two things to check.  Go to c:\users\\appdata\local\temp\vmware- and open the last file vmware-vmrc - xxxx.log.  Towards the end, it should tell you why it's a failure.  For me, it was a conflict of thumbprint SSL, dating back to the different SSL certificates on the cells.  Also, be sure to vCloud Director Administration-Public address you have the right audiences VCD console address specified. To resolve the inconsistency of my mark, I just created a file certificates.ks a cell and then copied on the other cell and reran the script configures.  Good luck with VMware to help.  I opened a folder for this issue 6 days ago and have yet to get any help!

  -Craig

• Without SSL on Apps Tier DNS load balancing

  Dear all,

  I am to set up load balancing DNS (non-shared file system) for my 2 nodes of level Apps. I followed the id of the note using Load-balance with Oracle E-Business Suite Release 12 [380489.1 ID] section 2.4, but the question is. my level Apps is not configured with SSL. Please tell me how to configure the DNS without SSL load balancing so that I can connect with http://

  
  

  Kind regards

  Aleem
  

  You can use the same instructions - replace https with http and 443 with your port number.

  HTH
  Srini

• Adpaters/active with the vPort ID load balancing

  Looking at the following configuration, my understanding is that some virtual machines in this group of ports will use vmnic1 and some will use vmnic6 (based on their ID vport).  Pretty simple.

  Assume that the uplink different swithces cards.  This works because there is no tree blocking one of the unplinks.  Question is: what happens to traffic on vmnic6 if the link fails? (failure of the local liaison, NIC failure, etc. - I'm even not consider upstream failure here).  My understanding is that vmnic1 will not resume because the windows are related to the vmnic6.  Only if there is a Standby adapter is there a link to relief for vmnic6 or vmnic1.  In other words, with vPort (no Etherchannel) of load balancing, two active/active adapters are not 'Standby' to each other.  Is that correct, or am I wrong?

  I ask because I see this routinely on customer sites.  But what I always have it set up as a channel of port for a single switch, if two switches or was to be used, then I made active/active and then included a Standby adapter for failover should one Fail of active principles.  Or, if all I need is resumed, not load balancing, I just did a single active/passive and alternate ports between them as a form of manual balancing groups.

  I always just assumed Active would not switch to another asset without a port channel/IP-hash config (where the whole reason for standby in the first place).  I'm quite right or completely wrong?  Later today, I want to test this is a lab while monitoring with top.

  

  With the default configuration in your example, the virtual machine-> Teddy assignment is when a virtual machine is running. In the case of a Teddy/link failure the VMs will be assigned to other natachasery. A standby NIC cannot be used unless there is a power failure.

  André

• TMS 14.5 and redundancy with F5 load balancing

  Hello

  I'm looking in the installation guide for TMS 14.5 to design redundancy TMS with the F5 for one of my clients. I wonder if there is any requirement of specific network for the TMS1, TMS2 and F5 must be in the same VLAN or not. My client have a TMS with TMSPE and deploy the TMS Server 2nd in different data center and will use existing F5 load balancing. They cannot have the same VLAN as a current TMS server in the new data where 2nd MSDS server will be Center. Existing F5 load balancing is also, in the VLAN different already.

  y at - there no specific review we need to take care of the redundancy?

  Thank you

  Chris

  As the TMS 14.4, TMS no longer an active/active load-balanced mode and now works as active/passive.

  You can find the details in the "Configure a redundant deployment" section of the Installation of TMS and Upgrade Guide (latest version 14.6 related).

  As long as both devices are accessible (and your F5 can route traffic to each of them according to the needs), it does not matter what VLAN, subnet or a location that each MSDS server resides on.  Information of VLAN in the guide are just one example of how it can be configured in a simple load balancing scenario.

  Wayne
  --
  Remember the frequency responses and mark your question as answered as appropriate.

• Load Balancing configuration failed with the failure of the connection shared after 11.1.2.2 11.1.2.3 upgrade

  Hello

  We have our EMP system had put 11.1.2.2 to 11.1.2.3 and it's a broadcast environment (Windows 2008 Server) and in one of the server, we installed FDM and after 11.1.2.2 upgrade to 11.1.2.3.

  When we load balance FDM Configuration, SSP is failed with " Shared Services Connection Failed '

  What I did:

  Deleted all entries in the object load balancing DCOM {E652643D...} and the Application Server Config {26AD6592..} and the COMJNIBridge DCOM objects. Re-directed EMP config for FDM application server

  and reconfigured the FDM with mycmsc\svc_hyperionfdm of the user application server. Config of load balancing is still unable to the connection of Shared Services with the error "Shared Services cannot connect".

  I tried the local administrator user but fails in the same way

  Windows system log entries:

  The application-specific permission settings do not grant Local Activation permission for the COM with CLSID {E652643D-6CC1-48AC-915D-01842B04F292} server application and the APPID

  {E652643D-6CC1-48AC-915D-01842B04F292} the user NT AUTHORITY\SYSTEM SID (S-1-5-1 of the address LocalHost (using LRPC).

  This security permission can be modified using the Component Services administrative tool.

  Windows Application log entries:

  Recovery of the class factory COM for component with CLSID {E652643D-6CC1-48AC-915D-01842B04F292} failed due to the following error: 80070005.

  COMJNIBridegWrapper.log entries

  08/07/2014 10:42:06: unknown COMException error

  at Microsoft.VisualBasic.CompilerServices.LateBinding.InternalLateCall (Object o, Type objType, String name, Object [] args, paramnames, CopyBack [] Boolean String, Boolean IgnoreReturn) at

  Microsoft.VisualBasic.CompilerServices.NewLateBinding.LateCall (Object Instance, Type, Type, String MemberName, Object [] Arguments, String [] ArgumentNames, TypeArguments, Boolean] [] Type

  CopyBack, Boolean IgnoreReturn) at Hyperion.Fdm.Api.Common.Security.Authentication.AuthenticateUser (String userName, String password, the domain, String String & ssoToken)

  Hello

  Finally, we could solve this problem. Thank you for all your help.

  Kind regards

  VJ

• For vRA 7.0 load balancer

  If one wants to set up vRA 7.0 as average deployment and use a load balancer, can configure us what with the NSX without another like F5 load balancer?

  Yes, it's OK - that ONE type of software load balancing is necessary.

  It is not that you need both F5 and NSX, one who will do it.

• NIC teaming + with leaves HP VirtualConnect and Cisco load balancing

  Hello

  I configure ESX infrastructure running on the blades HP (c7000 enclosure), VirtualConnect and Cisco 3750 switches for the uplink. My network configuration is based on the HP VirtualConnnect Cookbook, scenario 11.

  On the ESX Server, I configured a vSwitch with two NICs grouped. Each card can see a different network of VirtualConnect. Each VC network has active SmartLink.

  Each network of VirtualConnect is associated with two ports on a Bay of VC. These two ports are the uplinks in Cisco 3750 switches. Cisco switches are configured in a pile, and each port associated with the VC network is configured with LACP.

  Here are the facts:

  -Each network (Host A and host B) is properly configured for the LACP Protocol, in other words, the links are shown as active/active in the VC Manager.

  -Communication of ESX is good.

  -Failover and failback function OK.

  The problem: I can't get the load balancer to work. All virtual machines use a single bear. I tried different algorithms on the server ESX (source port ID, mac, IP hash hash) and the configuration of the equivalent from Cisco.

  I have attached a diagram of the physical network, but also the setting to ports.

  What Miss me? Thank you very much

  Pablo

  Because you want to create technically not a channel in an ESX perspective, you do want to use the hash of the IP.  If ESX attempts to send traffic through the two network cards, the cisco switch should drop this package as part of its loop avoidance algorithms.  You should be able to check that out in the switch port statistics.  You should not use the switch port ID.

  Not sure about the promiscuous mode.  Who do not have any account, as far as load balancing is concerned, unless the portgroup was not properly inherit the vSwitch properties?

  -KjB

  VMware vExpert

• Can we do vmotion between hosts with load balancing in network cards grouping different strategy.

  Hello

  We are implementin new host in our Infra and do some vmotions between different groups. A group a host with balancing 'route based on IP hash. " Can we do Vmotion to another cluster where the hosts have different load balancing policies. In addition, if we change the policy on new hosts in furture of load balancing is loss of downtime or a package?

  Kind regards

  Vikram Kumar

  First of all, no problem from VMs one ESX to another, they use different political Multipathing.

  Second, if you use IP hash, it means you are using etherchannel (or LACP) on the physical layer - configuration/cancellation of the configuration that almost always require that the uplinks are not used by any VMs, so you have time to synchronize the physical and virtual configuration. It is not a time out, but you need to plan accordingly - all depends on your physical network capacity.