Load balancing outbound requests from OAG
Hi,
I need to load balance outbound requests from OAG to a cluster of Application Servers (which expose some web services): do I need a hardware (or software) load balancer product? Isn't API Gateway able to do the job by himself?
Thanks.
Bye.
Roberto
Hi,
By adding a remote host you can achieve load balancing. So make up a "fake" host that should not exist just not to collide with any actual service, add that host as a remote host and then add the IP's in the list of the of load balancing. Us the fake host in your connection filter and the gateway will automatically replace it host with a correct IP from the load balancing list.
Similar Messages
-
Load balancing HTTP requests for an OC4J instance w/multiple JVMs
Hello everyone,
I am using OAS 10.1.3.1 and wish to load balance HTTP requests across an OC4J group of one or more OC4J instances, where each like named OC4J instance may have multiple JVMs or it may just have one JVM.
My mod_oc4j.conf file would contain the following directives :
Oc4jSelectMethod roundrobin:local
Oc4jRoutingMode Static
Oc4JMount /xyz/* xyz
In the degenerate case, I would like to have an OC4J group with a cluster size of one, and have that one OC4J instance have two or more JVMs. I would like to be able to receive a request within my web application, determine that the JVM that has been sent the request is the wrong JVM to process the request, and then call HttpServletResponse.sendRedirect("/xyz"). Then, when the browser receives the HTTP 302 response and issues the subsequent HTTP request, have that request sent to a different JVM than the previous JVM that issued the sendRedirect().
What I have seen is that the subsequent request is sent back to the same JVM that issued the sendRedirect(). I also call invalidate() against the HttpSession prior to calling sendRedirect(), but that does not seem to affect the behavior that I see.
In the full blown case, I would have an OC4J group with a cluster size > 1, and each of those OC4J instances would have at least one JVM. In that case, I wish the sendRedirect() call to allow the subsequent request to be sent to any one of the OC4J instances in the group, and any one of those JVMs within all of those OC4J instances.
Can anyone verify that my mod_oc4j mount directive is appropriate given the select method and routing mode? What else might I need to do to have a chance to have a different JVM respond to the request that results from a sendRedirect()?
Thank you,
DougI should clarify that in the full blown environment, the OC4J instances that form a group will each be housed within a separate OAS instance that resides on its own machine.
So ideally, a request could be inbound to say OAS instance 1 on machine A, OC4J instance AA, JVM 2, and I need to force a redirect so that the request can actually be serviced by OAS instance 3 on machine C, OC4J instance AA, JVM 1, and I need to be able to call sendRedirect() against an HttpServletResponse from within a JVM until the subsequent request from the browser, Internet Explorer in my case, is serviced by that JVM.
Thanks,
Doug -
Load Balancing HTTP requests to ABAP App Servers options?
Looking at SAP Documentations, SAP recommends to use the Web Dispatcher to load balance HTTP requests to multiply ABAP App servers.
My question is that the only solution? or can we use hardware such as the F5 BigIP to perform the same job?
Any thoughts?In collaboration with SAP and SAP customers, F5 Networks has created a solution that delivers security, high availability, and improved performance for SAP web and portal technologies.
By deploying F5 Networks solutions with SAP NetWeaver, enterprises extend their control over their Network and Application traffic, and ensure the fast and secure delivery of their applications.
Benefits of F5 for SAP NetWeaver and Enterprise SOA
u2022 Cuts SAP Enterprise Portal login time by more than half for WAN users
u2022 Speeds document downloads
u2022 Reduces SAP server CPU utilization by 44%
u2022 Provides a 20x reduction in the number of SAP server-side connections
Further details, case studies and deployment guides on;
http://www.f5.com/solutions/applications/sap/netweaver/
F5 certification information on SAP Website
http://www.sap.com/partners/directories/SoftwareISVSolutions.epx?context=21B87D61C0F646A22B2A6DB254A010CA8C9C141B7529F029910FE6FF9EEEC5A701BF20EED61AC07159D98BAA068EBE1B8C5C7665EA2226374E942CF1D2A49D20AB1BFDFA1E0B68EC41E3058F04A85F105D5002CF1A11383C905D9FE5DDB951251A4B574B0BBE58309F67667A3B95877FEF85F1EF8B2C1A9F6FBA3BF5066D9534%7c01518B8BD6BF02F55A5A72E5947F2C45
Hope this helps. Thanks -
I would like to use the CSS to load balance DNS requests for a telnet applicaion. I don't want my telnet sessions to go through the VIP, I want them to go directly to my host. I would like the CSS to run TCP health checks on port 23 and return the best DNS address based on least connections. I know you can do HTTP redirects, but I am not sure how to accomplish this with telnet traffic.
Thanks for the help, sample config would be appreciated.You could the information in this doc to configure health checks. But I am not too sure about the redirection and load balancing for telnet traffic. May be someone else could help.
http://www.cisco.com/warp/public/117/health_checks.html -
How to load a selective request from PSA to DSO using DTP?
Hi Gurus,
The PSA is having 3 requests, how to load the request number 2 from psa to DSO using DTP?
Thanks
SDP
Edited by: SDPSAP on Nov 15, 2011 7:58 PMLet us say that requests are 1,2 and 3. 1 being the first one and 3 is the last one. Now you need to load request 2 in DSO from PSA.
- create a Delta DTP
- Tick "Only Get Delta Once"
- Tick "Get One Request Only"
- Untick "Reterive Until No More New Data"
- execute the DTP, it will load request 1 to DSO
- delete this request from DTP
- execute DTP again, it will load request 2 to DSO; this is what you need.
If DSO is write optimized, you do not need to untick "Reterive Until no More New Data"; just execute the DTP, it will create 3 requests in DSO, delete first and last and you will get only 2 in DSO.
I hope it helps.
Regards,
Gaurav -
I get the process of load balancing outbound traffic, and even protocol binding so that servers send out traffic only on a specific WAN port. What I do not get is how to load balance INBOUND traffic also to those servers and ensure that the servers response goes out on the interface that it originated.
I have two 50mb internet connections from different ISPs. Each with a static IP address.
The web server is on a private IP address with appropriate forwarding.
The web server IP address is bound to WAN2 which is the DNS address of the web server.
All works perfect. Outboud traffic from the web server exits from the appropriate WAN port while residual traffic is load balanced.
Now, I want to add an A record to the DNS for the web server so that INBOUND traffic ot he web server is balanced and providing a redundant service. However, the protocol binding should not allow any outbound connection through WAN1. Right?
So my question is, is this setup impossible? Do we forget the protocol binding and hope that the outbound responses from the server exits through the originating port? Is there another method? Or, do I need a more advanced router?
I do not even want to consider DMZ even as this is impossible since I am utilizing both ports as WAN ports. Anyone have any suggestions to do this?
Thanks
MarkSorry I misunderstood.
IMO for this to work on an LRT224 you would need:
A second network adapter in the Server and the Server configured to answer requests on both NIC's
Protocol Binding the second Server NIC to the other WAN port
A DNS Hosting Service that support DNS Round Robin or you would run your own DNS Server with Round Robin DNS capabilities
These days I'm seeing DNS propagation with namesecure only taking a few hours instead of the advertised 96hrs.
Please remember to Kudo those that help you.
Linksys
Communities Technical Support -
WAN Load-Balancing and multi VLAN design
Hello,
I need some help to define the design of a specifi LAN-WAN network.
1) There are 2 independant WAN entries (they have their own ISP-managed router)
2) I need to load-balanced the requests over the 2 WAN
3) If possible, the load-balancer must be redundant (GLBP ?)
4) On the LAN itself, there must be 15 different VLAN
5) We also need a DHCP solution (also redundant if possible) to provide IP to these VLAN, with unique gateway (the load-balancer)
What do I need to implement this configuration ?
And is it possible to configure with as much GUI as possible ?
Thanks in advance for your help.Dear Mike,
Thank you and welcome to the Small Business Support Community.
It is possible to configure load balancing with NAT, however in this case, remote internet servers will potentially see sessions from remote hosts behind the SRP541W coming from different source IP addresses (the WAN IP addresses), causing the sessions to be reset unexpectedly.
The Policy Routing setting you setup is exactly what I would do in your case.
I hope these answer your question and please do not hesitate to reach me back if there is anything else I may assist you with.
Kind regards,
Jeffrey Rodriguez S. .:|:.:|:.
Cisco Customer Support Engineer
*Please rate the Post so other will know when an answer has been found. -
WebAS access via Portal: Web Dispatcher required for load balancing ABAP
Hi Folks -
We have EP 6.0 SP18 (Java only, WebAS 6.40, Unix/Solaris). The portal has a CI/SCS and one DI so we have a Web Dispatcher to load balance the portal servers. This works fine (and provides port 80 access).
This portal will provide access to HTTP services from an ABAP WebAS (6.20 with 6.40 kernel, Unix/Solaris). A landscape configuration entry has been added to the portal for this ABAP system. The ABAP system has a CI and multiple app servers, all capable of handling HTTP requests. This will also require port 80 access.
1. Will we need an additional Web Dispatcher to load balance HTTP requests to the 'backend' ABAP WebAS system, or will the portal be smart enough to handle the load balancing itself (perhaps based on the information in the landscape configuration)?
2. If the portal itself handles the HTTP load balancing can you point me to documentation (so I can make sure I have proper configuration)?
3. Are there any changes to this with NW2004s Portal (we plan to upgrade soon)?
Thanks in advance! JeffJeff,
Regarding:
Q1. If you create a system object from the "SAP system with load balancing" template in portal and configure the object to point to your CI (msg server), the LB should be handled.
Q2. Portal load balancing is handled by the message server. If you point a test URL to the port of your message server, you will notice that you are issued a redirect the URL of your dialog instance. The web dispatcher is just a proxy (with some intelligence). When a request is made to the WD, it makes a connection to the MSG server, the list of active instances is queried, a redirect is made to that instanct. If you use WD, that connection can be proxied behind a standard URL. If you connect directly to the MSG Server instead, you will notice your URL change, just as it does on the service marketplace.
WDs are good for providing services, masked (proxied) behind virtual names. If you do not want the customer to see a physical URL of the server, use the WD. There are lots of other solutions that can do this too though such as Apache, ISA, Juniper devices, Cisco LDs. WDs have a very low performance threshold though, especially if you use SSL. WD is a performance bottleneck and should be benchmarked to see if it is right for your application.
Q3. No changes this architecture in 04s.
jwise -
Hardware load balancing with session stickiness
Hello,
We are looking for a hardware based solution to load balance 4
unclustered weblogic servers. These servers are software replicas of
each other and share a common database. We need a solution that
provides session level server stickiness based on jsessionid as both a
cookie and url rewrite. Currently we have Cisco load balancers which
do not handle the url rewrite due to the fact the the load balancer
only "sees" what is after the question mark in the url and the
jsessionid is before the question mark. Example:
http://somehost.com/some/url.jsp;jsessionid=26D5C566075663ABD8E17AD255974323?parm=value
I'm sure we are not the only ones trying to do this. Any suggestions
would be greatly appreciated.
Thanks,
Alan George
Having worked on Cisco load balancer and WLS clustered, you would loose sessions
when the requests are failed over.
S
"Shawn Kircher" <[email protected]> wrote:
>
>Has anyone got this type of scenario working where you load balance sticky
>sessions
>from a Cisco load balancer between non-clustered weblogic servers?
>
>Thanks,
>
>Shawn
>
>"Aravind Krishnasamy" <[email protected]> wrote:
>>F5 or BigIp provides hardware level load balancing and it injects date
>>inside weblogic cookie and maintains server affinity.
>>Thanks
>>Aravind
>>"ageorge" <[email protected]> wrote in message
>>news:[email protected]...
>>> Hello,
>>>
>>> We are looking for a hardware based solution to load balance 4
>>> unclustered weblogic servers. These servers are software replicas
>>of
>>> each other and share a common database. We need a solution that
>>> provides session level server stickiness based on jsessionid as both
>>a
>>> cookie and url rewrite. Currently we have Cisco load balancers which
>>> do not handle the url rewrite due to the fact the the load balancer
>>> only "sees" what is after the question mark in the url and the
>>> jsessionid is before the question mark. Example:
>>>
>>>
>>http://somehost.com/some/url.jsp;jsessionid=26D5C566075663ABD8E17AD255974323
>>?parm=value
>>>
>>> I'm sure we are not the only ones trying to do this. Any suggestions
>>> would be greatly appreciated.
>>>
>>> Thanks,
>>>
>>> Alan George
>>
>>
>
-
SSL Setup in a load balanced portal
Hi,
We are implementing a portal landscape and also we are using a hardware based (Cisco ACE) load balancer for load balancing purposes.
So the configuration would be:
Portal requests --> Load Balancer --> Portal --> Backend
We are trying to implement SSL until the portal server and I have a question regarding the SSL certificate installation process.
The URL on the load balancer would be for example https://portaltest.mycompany.com which would load balance the requests between the application servers of the portal (https://sapeptest1.mycompany.com:50001/irj/portal and https://sapeptest2.mycompany.com:50001/irj/portal).
So, first thing we will have to do would be to install an SSL certificate (signed by a Trusted CA) on the load balancer with a CN=portaltest.mycompany.com.
I understand that for https to function properly, the host name in the URL we are using to get to the server should match the CN of the SSL certificate installed on the server.
Now, can we install the same certificate (that we put on the LB) on the portal as well?
(This might not work because the server type will be different)
(or)
Do we need to buy 2 certificates with the same CN and install one each on the LB and portal ?
Can some one please suggest on how to proceed with the SSL setup and certificate installation process ?
Thank You ,
RajRaj Kumar wrote:
My question is about how to go about installing the certificates on the LB and on the portal.
If you aren't using web dispatcher, then the details of the installation on the LB will depend on your LB (Cisco? Radware? etc?). I suggest contacting your LB vendor for that.
Sen's link is for SSO, you want the [SSL procedure|http://help.sap.com/saphelp_nw70/helpdata/en/f1/2de3be0382df45a398d3f9fb86a36a/frameset.htm].
You probably don't need a signed cert on the portal server itself (depending on whether your LB validates the cert). You could just use the default self-signed cert, since users won't be connecting to it directly and so won't be troubled by warnings about untrusted certs: the traffic from the AS would still be encrypted, you would only lose out on the server authentication feature (which you don't need, since again users won't see it).
On the other hand, do you really need SSL on portal server? That adds overhead at both the LB and portal. It's usually sufficient to use HTTP from the LB to the back-end, as long as the servers only allow connections from the LB. I realize you aren't using web dispatcher, but this looks like scenario #3 in [this diagram|http://help.sap.com/saphelp_nw70/helpdata/en/d8/a922d7f45f11d5996e00508b5d5211/frameset.htm]
Regards,
Sean -
short description of the setup:
2 hosts running report services say reportsvc_1, reportsvc_2 in the sanme subnet.
reports are called using run_report_object from within oracle forms.
Is there any mechanism that these two hosts can share the load, load balanced. I know run_report_object requires report service name in order to run.
what other mechanism can be used to load balance report requests initiated from oracle forms.
ThanxYou will need a virtual server between the app server and report server. Something like this:
App server
....|
....|
....^ - virtual IP server
.|.....|
R1..R2 - report servers
The virtual IP server accepts the call from the app server and rotates which 'back end' report server per call. It is a crude, round-robin solution, but it works relatively well. You have to hope that every-other report isn't rough on the server. -
HTTP logs not created when using software load balancer
This is my setup:
1. I have a server running a software load balancer - call it A
2. There are two servers clustered under A - call it X and Y
3. I am able to access the application deployed on X and Y through A.
Fail-over also works. Shutting down X when the application is up and running routes the requests to Y and vice-versa. All is well until this point.
Problem is:
Let us take server X.
$ORACLE_HOME/Apache/Apache/logs does not have an access log in it. (It
however has an error log.)
Access logs are not created when accessing the servers in the cluster (X and Y) through the load balancer URL.
However, when accessing the individual servers in the cluster (with their URLs), access logs are created.
Specifics on the setup:
1. A, X, Y - all three are OAS 10.1.3.2
2. All three are deployed on individual server machines
3. All three servers have one OC4J instance each running in them
4. The software load balancer used is from WebLogic 8.1sp5 - basically to create a *.war file that has the details of X and Y (servers in the cluster)
5. This *.war file is deployed on A
6. Server A, the one that has the software load balancer, DOES NOT have AS control. One of the other servers in the cluster does.
Has anybody encountered this problem before? Or is there anything wrong
with the setup? What is the resolution for the same?
Any help is greatly appreciated.
Thank you.Thank you, Jacco. You are right in saying that the software load balancer does not route the request to Apache and hence the logs are not created.
While the logs are not critical to the component we are working on, was just wondering why they were not being created.
As regards your comment on FastCGI, there is a FastCGI directory created under $ORACLE_HOME/Apache/Apache/logs - but there is nothing in it. Neither do the error logs have any information pertaining to FastCGI.
Anyway, I am not going to worry about this issue anymore - may be sometime in the future will try to use a different software load balancer and see if it makes any difference. -
User restrication on load balance VIP
we need to implement below requirement can any one suggest how to imlement it
1.Sticky command
Once a connection is opened to a physical server, any requests coming from a particular client always go to that server, until either the timeout is reached or the user session is terminated
2.Client-assigned load balancing
All requests coming from a specific client always go to the same physical server. This is done through recognition of the clients IP address.
3.User session restrictions
User name / IP session restriction based on parameter values(ie only 1 user can login at a time).
Please find the below config
service test1
ip address 10.8.1.25
protocol tcp
keepalive type http
keepalive port 80
active
service test2
ip address 10.8.1.26
protocol tcp
keepalive type http
keepalive port 80
active
content DSS-R1
protocol tcp
vip address 10.8.1.1
port 80
advanced-balance sticky-srcip-dstport
add service test1
add service test2
active
please suggest me how to implement all three points whereas point 3 is very crucial & urgent.1/ is done with the advanced-balance command.
I would suggest to replace the current one with 'advanced-balance sticky-srcip' since the destination port is always 80 [per config].
2/ you can force a client to go to a particular server with an acl and the prefer option.
ie:
acl 1
clause 10 permit tcp x.x.x.x destination content owener/DSS-R1 prefer test1
clause 99 permit any any destination any
apply all
acl enable
3/ we will need more details.
Not much can be done so, because the CSS only counts active connections.
So, you can restrict 1 user at a time with the command 'maxconn 1' but I'm not sure if it will help.
You should test it first and see if it does what you need.
Regards,
Gilles. -
Is it possible to load balance incoming requests based on client's operating system on ACE?
For example, we have different web pages specifically for Blackberry or iPhones.
Instead of having multiple URL's & VIP's, we'd like to have a single VIP, but load balance traffic to different serverfarms based on client's OS.You can loadbalance based on User-Agent header, first you need to quantify what Iphone and blackberry use for user-agent for instance from a regular browser you might see:
User-AgentT=Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 2.0.50727)
from an iphone you will typically see:
User-Agent=Mozilla/5.0 (iPhone; U; CPU like Mac OS X; en)
AppleWebKit/420+ (KHTML, like Gecko) Version/3.0 Mobile/1C25 Safari/419.3
you can go to http://www.user-agents.org to find out what strings are used
That being the case you can make classes on the header to match for loadbalancing decisions:
class-map type http loadbalance match-any mobile
2 match http header User-Agent header-value .*iphone
4 match http header Uswer-Agent header-value .*blackberry
then in LB policy say we want to go to farmA for mobile and farmB for pc's
policy-map type loadbalance first-match L7POLICY
class mobile
serverfarm farmA
class class-default
serverfarm farmB
see:
http://www.cisco.com/en/US/docs/interfaces_modules/services_modules/ace/v3.00_A2/configuration/slb/guide/classlb.html#wp1021388 -
HttpClusterServlet Load Balancing Problem
Hi,
I have a problem with LoadBalancing and HttpClusterServlet.
I have 2 boxes, with a clustered application, deployed on 2 managed servers
box 1 with adminServer and and a managed server (managed1)
box 2 with second managed server (managed2)
on box 1 I have another managed server, which hosts the HttpClusterServlet application
The cluster is made up of managed1 & managed2, and httpClusterServlet is configured
to send requests to boths the managed servers in the cluster.
When I run some tests against the httpClusterServlet web application to see how
it load-balances requests, I can see that all the requests are redirected just
to managed1,
on the same machine.
If I stop managed1, and run some tests, I can see that httpCluster every time
tries to contact managed1 first, and since it's down, sends the requests to managed2(handles
failover correctly).
If I move the httpClusterServlet application to another box, let's call it box
3, and re run the tests, it will load-balances the requests between managed1 (box1)
and managed2(box2) correctly.
I would like to know if httpClusterServlet has some kind of "server affinity"
logic (?) or I'm missing something in the cluster configuration (even if I can't
see clearly how the cluster configuration could affect things, since HttpServlet,
which is not in the cluster, just should route
requests to the specified servers in round-robin fashion, I guess)
Someone in previous posts pointed out http session issues, is it true? and so,
why?
Sorry for the far too long post,
thanks in advance,
Thomas
Only from load balancing point of view cluster is not really needed.
apart from session replication and failover capabilities, few more benefits are there with clustering only if you are using a web server also.
Stuck threads can be due to lot's of problems like application issues, load issues etc, So first make sure you are getting stuck thread on both servers or only on one, if on one, then in this case your load balancing not working properly and sending more load to that server, and if you are getting stuck threads on both then try to add one more managed server.
Regards
Mukesh Negi
http://weblogicserveradministration.blogspot.com
Maybe you are looking for
-
Adobe premiere elements 12 install with no disk drive
Please help me install adobe premiere elements 12 on my new computer with no disk drive. I appreciate your help!
-
I have a folder in Word 11 with word documents. Is there a way to access individual docments in icloud from my iphone or ipad without putting the entire folder on the iphone or ipad?
-
I want to buy an external keyboard for use when I'm not traveling. I am considering the Apple Wireless Pro keyboard and would welcome any opinions about that one or any others (wireless OR wired) you like or dislike. Another question is about mice/tr
-
HT1918 have a new debit card try and edit paymeny info but keeps saying its expired but its new
I have a new debit card, trying to edit payment info but it keeps saying its expired although its a new card.
-
Is it possible write code inside pdf form when click submit button in pdf form
can write code to pdf form for eg: we have to write code in Excel sheet using Macro features Like that any possible to write code in pdf form when i downloaded pdf form from site then filled up the details then click the save button the pdf form will