LACP ports used as autonomous ports

Similar Questions

• Use of ISE 3355 device of the two GigE ports

  I'll put in place six ISE 3355 3 devices in a data center in another 3. They just installed a new infrastructure of server using 5596 Nexus and Nexus 2248TP farm top of rack switches.

  I am looking for documentation on how to make the collection of NETWORK adapters on the way 3355 or some connect Gig0 to FEX101 and Gig1 to FEX102. Or simply set up a channel using LaCP port between the two different groups of FEX?

  Sent by Cisco Support technique iPhone App

  Hello

  This is not supported, you can not team or balance the load or use redundant interfaces on devices of the ISE. You can only use a span port dedicated for ISE deployments, or use the links to crossover to a deployment ipep in HA mode.

  Thank you

  Tarik Admani
  * Please note the useful messages *.

• 3650 LACP etherchannel suspended

  This new series of 3650 has been bugs me for a year. A lot of things that used to work is not now.

  Over the last weekend, I upgraded a 3560 X to a 3650. The etherchannel seems to behave differently in 3650.

  Example,

  Interface Port-channel10
  switchport trunk allowed vlan 17.23
  switchport mode trunk

  interface GigabitEthernet1/0/11
  switchport trunk allowed vlan 17.23
  switchport mode trunk
  active in mode channel-group 10

  interface GigabitEthernet1/0/12
  switchport trunk allowed vlan 17.23
  switchport mode trunk
  active in mode channel-group 10

  Now when this config is used in a 3560 X and the terminal is in mode standby (not to communicate LACP) ports go into I - autonomous State.

  The same configuration is now placed in a same scenario 3650, and ports through s - suspended state.

  This may not seem to be a big problem, until I try to send WOL for the terminal. Autonomous State, he will answer, but in the s state effectively the line protocol is down the WOL packet (suspended), cannot get there.

  I was really disappointed after a year through versions 3.3.3 to 3.7.1 there is still funny issues with the 3650 series.

  I hope someone is able to advise if they have a simple fix for this problem.

  Hello

  Add the command no channel of port authority - disable the port-channel interface and this should do what you want for example, interface member will in particular (I) rather than the State Suspended (s) if it does not PDULACP. You may have to bounce the port-channel interface for that to be taken into account.

  Take a read of the command in the section configuration LACP autonomous or independent Mode of the configuration for more details guide.

  Concerning

• LACP configurations 6509

  Hello everyone, I looked around the community to solve a problem on creating connections between my 6509 Ehter channel and a server using Broadcom Advanced Control Suite 4 to create the collection of NETWORK adapters.

  I found different configuration settings that revolve around similar settings, but in some cases items are excluded or included I may need.

  Our facility is 2 servers to connect to 6509 s 2.

  Each 6509 a g9/17, g9/19, g9/21 and g9/23 ports allocated for use.

  The idea is to have spend A connect on the first 2 and B connect to the second 2 on both 6509 s.

  9/17 and 9/19 ports must be configured for LACP create channels of ether.

  Someone at - it a good document referred me so I can create configurations?

  So far, I have:

  int g9/17

  no ip address

  channel-protocol lacp

  active in mode channel-group 1

  LACP port-priority 1

  I've seen other configurations which create the channel protocol int first, then trunking, which we don't think that we need.

  The addition of VLAN native plus the server VLAN which is located in the server.

  I really need to know when to use which type of configuration.

  EJ

  Normally configure you ports members first and make sure they have a compatible configuration. If you want to run it in L2 and with a single mode VLAN then set the Member access and in the correct VIRTUAL LAN ports. Something like:

  conf t

  int rank gi9/17, gi9/19

  switchport

  switchport mode access

  switchport access vlan x

  active in mode channel-group 1

  channel-protocol lacp

  Daniel Dib
  CCIE #37149

• 5.1 current performance but LACP options not available

  I read in the article vmware knowledge base below to discover how configure LACP on a dvswitch I understand is supported under vsphere 5.1

  http://KB.VMware.com/selfservice/microsites/search.do?language=en_US & cmd = displayKC & externalId = 2034277

  The link above in the section "activating or deactivating LACP we year Uplink Port Group" all right until I get to the point 4.  "In the LACP section, use the drop-down menu to enable or disable the LACP Protocol."

  I can't see this section?

  The only thing I can think is that this is a problem with the fact the dvSwitch I'm trying to configure was created when the esxi and vsphere hosts were running on version 5.0 Update 1.  Now that I just upgraded to 5.1 on all hosts of ESXi, vCenter etc it is possible, that this feature has not been applied to this existing dvSwitch?  Will it create and configure a new dvSwitch?

  Any help appreciated.  Ideally, I would like to not have to recreate the dVSwitch didn't because he has a lot of configuration went inside.

  

• Output redirection autonomous Agent

  Hi all
  
  I use the autonomous Agent on the machine of DB and I'm trying to redirect its output to a file. I tried the following:
  
  ./agent.sh-NAME = cloudDB_agent-PORT = 20910 - v = 3 & > agent.out 2 > agent.err
  
  but I'm still getting:
  2012-09-25 12:02:31.868 Agent of NOTIFICATION ODI-1128 cloudDB_agent begins. Application server: AUTONOMOUS. Agent version: 11.1.1.6.0 - 27/02/2012. Port: 20910. JMX port: 21910.
  2012-09-25 12:02:37.624 ODI-1136 NOTIFICATION from planners on Agent cloudDB_agent.
  2012-09-25 12:02:38.329 Agent of NOTIFICATION ODI-1111 cloudDB_agent began. Agent version: 11.1.1.6.0 - 27/02/2012. Port: 20910. JMX port: 21910.
  2012-09-25 12:02:38.428 NOTIFICATION ODI-1137 Planner started for referential work WORKREP1 on cloudDB_agent of the Agent.
  ...
  
  printed on my terminal window, which is annoying.
  
  Am I missing something? I should start the agent by using another script?
  
  Thanks in advance.
  
  Joan

  Joan...

  This is the output at the start of the ODI Agent... :)) I don't think you can avoid it...

  You can try this

  ./agent.sh-NAME = cloudDB_agent-PORT = 20910 - v = 3 & > agent.out 2 > agent.err &

  that is to put an ampersand at the end... to run as a background task... You will see it not throw any message :)

  It may be useful

• autonomous transaction block of exception

  Hi friends,
  
  Can you tell me if it is possible to use an autonomous transaction block exception?
  
  If so, how? Please provide me with a guide line
  
  thnx

  ROY wrote:
  Hi friends,

  Can you tell me if it is possible to use an autonomous transaction block exception?

  If so, how? Please provide me with a guide line

  You cannot use an autonomous transaction directly in the exception block:

  SQL> ed
  Wrote file afiedt.buf
  
    1  declare
    2    v_dummy number;
    3  begin
    4    select a into v_dummy from testtab where a = -1;
    5  exception
    6    when no_data_found then
    7      declare
    8        pragma autonomous_transaction;
    9      begin
   10        insert into testtab (a) values (-1);
   11      end;
   12      raise;
   13* end;
  SQL> /
        pragma autonomous_transaction;
               *
  ERROR at line 8:
  ORA-06550: line 8, column 14:
  PLS-00710: PRAGMA AUTONOMOUS_TRANSACTION cannot be specified here
  

  but, if it is declared in a procedure you can call it the exceptional block...

  SQL> ed
  Wrote file afiedt.buf
  
    1  declare
    2    v_dummy number;
    3    procedure insertit is
    4      pragma autonomous_transaction;
    5    begin
    6      insert into testtab (a) values (-1);
    7      commit;
    8    end;
    9  begin
   10    select a into v_dummy from testtab where a = -1;
   11  exception
   12    when no_data_found then
   13      insertit;
   14      raise;
   15* end;
  SQL> /
  declare
  *
  ERROR at line 1:
  ORA-01403: no data found
  ORA-06512: at line 14
  

  or

  SQL> ed
  Wrote file afiedt.buf
  
    1  declare
    2    v_dummy number;
    3  begin
    4    select a into v_dummy from testtab where a = -1;
    5  exception
    6    when no_data_found then
    7      declare
    8        procedure insertit is
    9          pragma autonomous_transaction;
   10        begin
   11          insert into testtab (a) values (-1);
   12          commit;
   13        end;
   14      begin
   15        insertit;
   16      end;
   17      raise;
   18* end;
  SQL> /
  declare
  *
  ERROR at line 1:
  ORA-01403: no data found
  ORA-06512: at line 17
  

• Autonomous

  can give everything
  Autonomous_transaction
  
  with pragma
  
  
  Notes:

  I hope to read the articles you want to conclude the following:
  There is only one "legal" reason to use an autonomous transaction: in a logging error routine.
  You must avoid the autononous transactions in all other situations like triggers etc.
  http://asktom.Oracle.com/pls/asktom/f?p=100:11:0:P11_QUESTION_ID:1171766400346817259

• How can we get the 3.6 < preload > element used?

  Coherence 3.6 adds a new item * < preload > * the configuration of the local system. 3.6 developer guide documents this element as follows:
  
  + "Indicates whether or not a cache will preload the data of its object CacheLoader (or dumps).» Valid values are true and false. Default value is false. » +
  
  I couldn't find no other mention of this element in the documentation.
  
  Can anyone describe how this configuration element is intended to be used?
  
  Phil

  Hi Phil,

  Yes, you are right - it's IterableCacheLoader, not store, sorry about that.

  I'm not sure of the config element, but I'm sure I've seen documentation on the LocalCache.preload () method before, and this element has been around for a while (doc'ed or not - I guess it was a well-kept secret ;-))

  Part of the reason may be that it sounds like it will allow you to pre load your caches (this is what I thought the first time I saw him, until I dug deeper into the code and discovered what he does). Problem is that you can use for autonomous local caches, and even in this case, the way that he don't preload is not quite optimal for a typical scenario of loading data in an RDBMS.

  That said, it might be useful in some situations (to load small sets of reference on each node read-only data, for example), so without a doubt, I agree that it should be better documented. I hope that someone on the team of doc follows this thread ;-)

  See you soon,.
  ALEKS

• What switch do you recommend

  Hello

  Brief description of our environment...

  x 1 poweredge R720 with x 4 network ports for iSCSI
  x 2 poweregde R710 with x 4 network ports for iSCSI
  x 1 poweregde R610 with x 2 network ports for iSCSI

  x 2 PS6100E EQL
  x 2 PS4100E EQL
  x 2 PS6000 EQL
  X 4 PS5000 EQL

  Everything is distributed evenly between two floors of the building by using SyncRep and regular replication on the EQL boxes and using a VMware HA and DRS cluster on hosts.
  Physically separate networks for iSCSI network and production.

  We use 5448 powerconnect on the iSCSI network switches. One PC5448 for each floor and they are connected with x 4 fiber LACP trunk link.

  I know that these switches have to be replaced, but what? PC6248, PC7048 or Force10? Perhaps even another manufacturer? If so which model? The price IS a decisive factor, I don't have the budget for expensive stuff

  Check the support matrix Equallogic to get an idea of what switches are validated and supported / recommended. It is not limited only the Dell switches (Cisco, Juniper and Brocade options listed there).

  First, you'll want to upgrade to double switches on each floor and LACP together. You can stack them, but this has the drawback that if ever you want to / need to upgrade the firmware of the switch in one or both of the rooms, you'll need 100% service interruptions in this room.

  In my next part I'll stick to switches Dell since it is a Dell message board.

  I would skip the PC6200 series unless your budget can not install a switch better. The 6200 is a decent switch, but it's on its way out (this can be seen by the fact that Dell has not updated the firmware on the switch since the 3.3 series (all 7 k and higher switches are now on a firmware 5.x).) The Force10 S25 series is better than this PC7024 a switch, but if you plan to switch to 10 GB at some point, you may consider a pair of 8132 s PowerConnect in each room. The non - F version uses RJ45 ports that can make 1Gbit and 10 GB, then you can use to start just the current units of EQL, but new units could be 10Gbit units. You want to get the QSFP + optional modules for each switch and 2 wires for each pair, so you can use these 2 40Gbit LACP ports switches set in each room.

  If you want to use a LACP between the 2 rooms consisting of two switches, you have a more difficult choice to make; stack the switches (with the disadvantage that I mentioned above), or upgrade to Force10 S4820s that these offer the possibility to make a channel of port on 2 non-Stack switches (PowerConnect firmware does not offer this).

  Cisco and Brocade offer similar characteristics (to execute a LACP on 2 switches non-Stack), but I do not know what exact firmware and/or switches that would offer.

• switch the product selection, selection based on feature requirements

  Ladies and gentlemen,

  I'm looking for switches from linksys for awhile. and I used to work type cisco 3750 and 6500 Series.

  but now I want to achieve these features in a type of smb linksys/cisco switch

  -a switch with two battery units

  -lacp/lag a portchannel to shared resources with two ports, with a port to each switch

  and ofcource the abiliti to turn off a switch if necessary / recover form a power failure in a powergroup. (ofcource or a failure of the unit) without service interruption.

  support are a stack sharing all resources as would a pile of cisco 3750.

  When I read the cheat sheet the SGE2000, SGE2010 are stackable, but they can do the same job that a 3750 would be?

  I read some strange things on the forums with the creation of shared resources lacp ports if this already gives a fight that I have a hardtime assuming that it also works with an lacp in a switch stacked with a single cable in each unit.

  does anyone have experience with this?

  should which model I?

  Thanks in advance

  Soul

  As I understand it, we support gal cross stack on the EMS series, although I have not tested it myself.  I'm curious to see Christopher tests or confirmation as well.

  W / the 3750 s, the whole stack will appear as a switch, even with the SEMs.  With the case of the 3750 s, they have a dedicated interface of stacking w / built in redundancy.  The EMS series uses Gig interface ports.

  You mention only two of these switches, so communication between the battery should not be too large.  Also to note, most traffic normally of the edge in the soul, so for most types of application, this means that most of the traffic will cross the cross-stack OFFSET or the uplink.  It is ok.

  Multicast comes to mind though... and you have a lot of traffic mcast, then the 3750 s and pile it higher throughput would be the ideal way to go.  You don't want a lot of traffic replicated in 1 GB stack that the EMS would be a part of...

  Any thoughts on if you SmartNet need?

  I'm partial to the 3750 s myself (as you can probably :-)), especially because you can do almost anything with them.  Very powerful and flexible to integrate and fix any problem. A nice switch.  You are considering also the 3560 series or another product for these installs?

  I hope these comments help, greetings,

  Andrew

• How to create Spatial indexes in SQL Developer Data Modeler 4

  Hello

  What is the procedure to create a Spatial index in SDDM v4? I found myself at an impasse because of the following problems:

  (1) retro-engineering rate the spatial index, then I need to add them manually.

  (2) adding an index does not allow me to check the 'spatial', and does not allow me to add the sdo_geometry column in the index definition.

  (3) for the registration of the spatial properties for a table, I can not choose the name of the Spatial Index.

  I came across an old post (Data Modeler - cannot create a spatial index), but this post is no longer valid because SDDM V4 does not allow me to create an index without columns (definition of the incomplete index).

  Help is appreciated,

  Kind regards

  Richard.

  Hi Richard,

  Thanks for your update.  I'm scared, I was assuming you were using the stand alone rather than SQL Developer Data Modeler.

  I tried using SQL Developer 4.1.0.18.37 (ai2) using a thin connection and got the ClassCastException you have found.

  But when I used ai2 autonomous, maker of data using the same login details 'Advanced', it worked OK and imported from the spatial index.

  (I'm puzzled as to why these two seem to behave differently, as this is the version ai2 data integrated with SQL Developer ai2 maker).

  I then plugged and reverse engineering the same tables (yet) in a new model.

  Funny is that there is no error in the external journal this time, but the spatial properties are still not retroconcues correctly.

  Is it possible that you have not set the option to import the spatial properties?  I expect that the spatial properties would be imported or that the error should be registered.

  SQL Developer ai2 imported properly the spatial properties and the spatial index, when I used a basic connection and specified host, port, and SID or Service name.

  Another issue: SQL Developer relies on a certain JDK / JDBC library?

  Any Java JDK 1.8 should be good.

  David

• ESXi 4.0 / 4.1 Broadcom iSCSI installation program / Support with unloading iSCSI

  Hello.

  I am new to the use of iSCSI on ESXi so I may have a few questions already answered. Sorry for that.

  Is it true that ESXi 4.1 supports iSCSI unloading, while ESXi 4.0 does not work?

  How to set up properly? I have a Dell R510 with an Quadport Broadcom NetXtreme II 5709.

  I like to use two network of 5709 with links LACP ports to connect to my iSCSI target (QNAP).

  I found a few messages about creating two vSwitches which each card of a network and mapping the vSwitch (vmkx) to the vmhbaXX (esxcli swiscsi nic add - n vmk1 vmhba35 - d).

  What is the right way? How to set up link aggregation then?

  This is what it looks like:

  

  

  Thank you

  Michael

  Hello

  you look here: http://blog.open-e.com/bonding-versus-mpio-explained/

• ODI physical Architecture Agent not working

  Hello
  
  My ODI server is on Linux and the client is windows. I'm testing the agent and I get the following error.
  "ODI-1424 - Agent Host or port can be reached using http://localhost:20910 / oraclediagent".
  
  I have the following values
  
  Host = localhost
  Port = 20910
  Content of the Web Application = oraclediagent
  Protocol = http
  Maximum number of sessions = 1000
  
  Please let me know how to solve this problem. Thank you for yor time and help.

  autonomous agent doesn't have to start adminserver/managedserver weblogic, weblogic / a is used for the j2ee agent, odiconsole and public webservice.

  your independent agent can be installed in any machine, not always on the same machine as weblogic or customer.

  If your client's Studio, and it is different from the machine that autonomous agent is started, when to set physical agent, do not use localhost, you must use the autonomous agent machine name or ip address.

  BTW, before trying to start the agent of standlaone, you must change odiparams.sh or .bat to provide correct information.

• Grouping of active/active network adapters

  I have two 6509 cisco switch

  I have configured the two physical network adapters on ESX 4.1 as active/active and they are connected to a NIC in each switch.

  That way I have more flow?

  How ESX handles the traffic in the pool of active/active NETWORK adapters?

  NickHorton wrote:

  As I understand it, (even once that caveat, not a network) you cannot extend a vswitch on switch ports without a violation of the spanning tree protocols that disables anything but a port UNLESS your links aggregation switches battery cross. Someone correct me if I'm wrong, please.

  You are right. There is some confusion of terms only.

  There is a network standard called 802.3ad (working group name) or 802.1AX (real name now, but the name sticks around the working group.) This explains how two units may group several links together and logically to treat them as one. On Cisco switches, this is called an 'association' on the HP Procurve switches, this is called a 'trunk', and on most servers, this is called a "NIC team." On ESX, it is called "IP hash.

  Both parties must be configured to do so, otherwise the same MAC address is displayed both on different ports and this would cause various problems, including trouble of Spanning Tree, but also other things.

  The 802.3ad has two modes, a static configuration, or dynamics of the aggregation of links. The static method is quite simple, the two sides should only be configured manually. When you use dynamic a certain protocol called LACP is used to negotiate it. Cisco also has its own propertiary Protocol PaGP that does the same thing.

  The host ESX/ESXi only has support for aggregation of static links, so this must also be set on the side of the physical switch.

  The 802.3ad/802.1AX does not specify HOW loads should be balanced across the available links. It is up to the different units to decide and it could very well be different on two sides of it without problem.

  Various switches (physics and also the Vmware vSwitch) uses a method that includes IP addresses in the source and the destination to choose a NIC out this load balancing works well and is quite simple to do, but the boundaries a session between two periods of investigation to a single network adapter. If the connection fails, it will go to another.

  Some switches have the possibiliy to use more advanced algorithms to do this, including for example src/dst IP and also the TCP port numbers. It is more "expensive" to do what the package should be inspected deeper and also cause more general CPU, but will allow the load between two IP hosts use multiple network interface cards. It is not supported on the standard vSwitch or the distributed switch, but on the 3rd party Nexus 1000v.

  However, all methods is the aggregation of links, some differences in how to place images on the network adapters.

  As for using two different physical switches standard does not allow this. Aggregated links must begin and end between the two groups. However, several switch suppliers (for example, Cisco and HP) formed their own solutions for this, called stacked Switch Etherchannel or Trunking distributed for example.

  IF you have these switches with this function and if set up correctly on the side of the switch, you can for example set four vmnic "hash IP" on your ESX/ESXi and connect these two physical switches. This would probably be the best combination of fault tolerance and load balancing.