Verity Issue - Clustered Servers

To summarize:
We have an interesting issue that I am attempting to create a
work around for. Our website exists in a clustered enviorment (3
servers), a development server, and a file server (mostly PDF's
Word docs, PPT's etc...). Each server re-indexes the Verity
collections nightly via scheduled task and the file server does as
well. Since we host our servers in a sercure external enviorment
there is no direct network connection between the clustered servers
and the file server. Our work around for this was, since our
development server can see both, we would:
1) Re-index nightly on the file server
2) Set a scheduled task on the dev server to copy the files
from the collection on the file server to the test enviorment
3) Use our source control to submit these copies to the
clustered servers via scheduled task.
Everything was working great until we upgraded the file
server to MX 7. The collection file structure has changed for
vertity in MX 7 and we cannot output the results on the search
pages. We do plan to upgrade all of our servers to MX 7 but it is
going to take a few months.
We are trying to come up with some ideas as how we can
integrate the results from the file server. (i.e CFHTTP) as we dump
the results from 3 collections into an array and then sort the
relevance.
Any thoughts?
Thanks,
Jon

I'll restate this in simpler fashion.
Can you retrieve MX 7 verity search results on a MX 6.1
server?
Can this be done via CFHTTP or any other similar method?
Thanks

Similar Messages

Running BizTalk on multiple non-clustered servers. Potential serious issue

Hi,
I have just discovered what has potential to be a fairly large issue in our current BizTalk Architecture. We are running BizTalk Server 2006 R2 on 2 different servers, running Windows Server 2003 R2. These servers are NOT clustered. Instead, BizTalk is installed
to both, configured on 1 and then on the 2nd, they have joined the existing BizTalk environment. We have 1 host with 2 host-instances for say an Orchestration Host. (So 1 Orchestration Host and then 1 Orchestration Host Instance on each server), we also have
a host instance for Receive and Send on each server.
This is a problem that I have identified on an application currently in UAT:
We have a routine called 'Recovery Mode'. It doesn't matter what Recovery Mode is but what does matter is that only a single instance can go into recovery mode at the same time.
Consider this:
bool inRecovery;
while (isBroken) {
if (!inRecovery){
inRecovery = true;
// Recover
inRecovery = false;
else {
// Wait, recheck until not in recovery, or fixed.
Imagine an instance on server 1 needs to go into recovery, it first checks the inRecovery flag, gets returned false so goes into recovery, let's say this takes 5 minutes to complete. During this time, an instance on server 2 needs to go into recovery, checks
the flag, but the flag on server 2 has not been set to true.... Now I have 2 instances in Recovery...
My questions are:
1. Can I cluster 2 non-clustered servers? I'd need to upgrade from Windows Server 2003 R2 Standard to Enterprise is that possible? If so, Can I simply upgrade and cluster the servers or would I need to format the disks?
2. If no to number 1, where would you store the 'inRecovery' flag'? I can't store in a DB as there will be slight cross over where instance 2 might return false in my example just as instance 1 is writing unless I can lock it on write or something.
Some advice would be greatly appreciated.
Thanks
Rod

To answer your specific question:
1. Yes, but you would need to upgrade to at least Enterprise. Windows Server 2003 R2 Standard does not support clustering. IIRC, the upgrade from Standard to Enterprise is pretty easy, perhaps just a SKU change.
2. If you absolutely cannot upgrade Windows Server, then you'd have to store the state externally, such as SQL Server. There shouldn't be any concurrency issues if you do the check in the right sequence. Basically, begin a transaction, attempt
to UPDATE, return weather the UPDATE was successful, meaning that call grabbed the lock. If you check then update, yes, there is a miniscule chance both will return the same value but only one near-sumultaneous UPDATE will succeed.

Tlisten resolve incorrect IP add on Clustered Servers

Hi All,
I am having a problem in deploying Tuxedo v6.5, as a stand alone application, on
a pair of MS NT 4 clustered servers. The problem is when starting tlisten process
using standard command "tlisten -l //HOSTNAME:12345", tlisten picks up a different
IP addresses every time I start another tlisten process.
i.e. starting three tlisten processes may result in three tlisten processes listening
on three different IP addresses that belongs to the cluster pairs.
My two constrains are:
(1) I need 6 tlistens processes to be started on the same physical machines
(2) my client insist to use Hostname unless it is proven that it is a defacts in
BEA's tlisten
I have investigate with the Network support ppl, and according to them, their DNS
settings seems to be fine (i.e. they suggest that they have only map HOSTNAME to
a single IP address).
Any help would be appreciated.
Regards
Nathan

I added the program under the allowed apps but it still didn't work.
I disabled the firewall but it didn't work either.
Added IP addresses under DNS somewhere, but didn't work.
This would only handle the built-in software firewall that is part of the OS X operating system on your Mac.
I was told I may need to configure access to port numbers and/or server addresses. Certain ports should be open and certain server addresses "added", and it is preferable to configure a firewall by server address instead of IP address. If an IP range exception can be entered, I will need to enter a range of about 100.
By default, all incoming traffic from the Internet is blocked by your router. On the other hand, all solicitated outbound traffic is NOT blocked. Note, however, that the Apple routers do NOT support UPnP. If your application is coded to do so this may be the issue you are seeing. Using a non-Apple router would be one solution.
Opening ports on the Apple router is done via the AirPort Utility. Please check out the following AirPort User Tip for details.

TED with Clustered Servers

Hi All,
I am wondering what I need to do or even if TED can, set a package path to a clustered volume path instead of setting the package path of a distributed application to a single file server name?
TED itself is not cluster aware but can it be tricked/manipulated?
eg.
Currently apps come down and the package path changes from:
\\EnterpriseDistributorServer1\APPS1\TheApplicatio nName
To:
\\LocalFileServer1\APPS1\TheApplicationName
As it should and always has with single server sites, however our clustered SAN connected environment is also doing this. Which doesn't take advantage of the clustered servers resilience capabilities at all and renders applications unavailable if LocalFileServer1 goes down.
Instead I want it to go from:
\\EnterpriseDistributorServer1\APPS1\TheApplicatio nName
To:
\\Cluster1\APPS1\TheApplicationName
I want to do this without setting Cluster1 as the package path at the Enterprise level and configuring the distribution to keep the same source path, so our other servers maintain their own versions locally when distributed.
I also don't want to have to create secondary NAL and distribution objects just for the cluster, as I also receive enterprise distributions that I do not control apart from subscribing to.
Any help/ideas would be much appreciated.
Cheers,
Brad

brad,
It appears that in the past few days you have not received a response to your
posting. That concerns us, and has triggered this automated reply.
Has your problem been resolved? If not, you might try one of the following options:
- Visit http://www.novell.com/support and search the knowledgebase and/or check all
the other self support options and support programs available.
- You could also try posting your message again. Make sure it is posted in the
correct newsgroup. (http://forums.novell.com)
Be sure to read the forum FAQ about what to expect in the way of responses:
http://forums.novell.com/faq.php
If this is a reply to a duplicate posting, please ignore and accept our apologies
and rest assured we will issue a stern reprimand to our posting bot.
Good luck!
Your Novell Forums Team
http://forums.novell.com

Can I use Admin Console to Start Clustered Servers ?

          I have just started to look at clustering and am in the process of running through
          the cluster example tutorial. I have used the domain config wizard to create
          a manged server with a cluster and another server to act as the http proxy. As
          I prefer to use the admin console for all tasks I wondered if there is any way
          to start the managed servers from the console. The reason I ask is that I cannot
          find a way to start another server from the console apart from the admin server.
          On a seperate note can anyone tell me whether the multicast IP address is something
          that I should be concerned with as a developer. The reason I ask is that I know
          it is used for clustered servers to communicate, but how can it do that if the
          IP address is a physical box that does not have BEA installed.
          Thanks
          Justin


Yes, you can start all managed servers from the console ... Look under the
          cluser that you create and click on the control tab ... you should be able
          to start all the managed servers that will be apart of the cluster there.
          Weblogic uses a service on the local machine called node manager which will
          start and stop the services on each managed server.
          Its good for developers to know something about multicasting as far as what
          it does, how it works but most system administrators all well versed with
          the inner workings of it. You have to have weblogic installed in order for
          that particular box to become apart of the cluster, and you can run multiple
          managed servers on one physical box but be careful with resource intensive
          apps.
          "Justin Flanagan" <[email protected]> wrote in message
          news:[email protected]...
          >
          > I have just started to look at clustering and am in the process of running
          through
          > the cluster example tutorial. I have used the domain config wizard to
          create
          > a manged server with a cluster and another server to act as the http
          proxy. As
          > I prefer to use the admin console for all tasks I wondered if there is any
          way
          > to start the managed servers from the console. The reason I ask is that I
          cannot
          > find a way to start another server from the console apart from the admin
          server.
          >
          > On a seperate note can anyone tell me whether the multicast IP address is
          something
          > that I should be concerned with as a developer. The reason I ask is that
          I know
          > it is used for clustered servers to communicate, but how can it do that if
          the
          > IP address is a physical box that does not have BEA installed.
          >
          > Thanks
          >
          > Justin
          Outgoing mail is certified Virus Free.
          Checked by AVG anti-virus system (http://www.grisoft.com).
          Version: 6.0.638 / Virus Database: 409 - Release Date: 3/21/2004

Set logging severity level for clustered servers

How can I set the logging severity level for clustered servers just once at the cluster level rather than one at a time at the server level?

I know the logging is server by server, but surely there is a way to configure the multiple servers with one change rather than one by one. We have 40 some servers in the cluster.

Verity and Clustered Coldfusion Servers

Hi All,
I am currently looking at using Verity with our clustered
coldfusion servers but seem to have an issue. We currently have
three coldfusion servers but the version of verity included with
coldfusion only allows a connection from one server.
Has anyone else been in the same situation where they need
more than one coldfusion server talking to the same verity server?
If so how did you solve this issue? ( We have looked at purchasing
the full version of verity but it is to expensive.)
Thanks
Paul

I'll restate this in simpler fashion.
Can you retrieve MX 7 verity search results on a MX 6.1
server?
Can this be done via CFHTTP or any other similar method?
Thanks

Sharepoint 2013 RBS (remote blob storage) issue - Clustered datatbase

I have configured and enabled RBS on the CLUSTERED content database as per the following article
http://technet.microsoft.com/en-us/library/ee748631(v=office.15).aspx#enableRBS
However, I need to clarify the following as it is different from an article to another: 1- Do I have to install the RBS.msi file on the sql server? or just the web servers in the farm? 2- After installing and configuring everything as per the article mentioned
above, sharepoint still stores documents in the content database and not in the blob folder. Anyone has any clue on how to troubleshoot this?
Thanks and your help is much appreciated

Hi omirza,
Thank you for your question.
We are currently looking into this issue and will give you an update as soon as possible.
Thank you for your understanding and support.
Best Regards,
Eric
Eric Tao
TechNet Community Support

ITS Login screen behavior different on 2 clustered servers(BBP)-config same

We have a production server(s) for Standalone ITS 6.2 (BBP) with 2 hosts and this is load-balanced and clustered.
The templates for (BBP) bbpstart and bbpglobal for the affected instance are identical on both physical servers.
The services on both the servers are also identical. But yet, Login behaviour on the 2 servers are different.
The servers are :-
Prd:    (Instance) WEP     on server it03 and it04 port 82
URL: http://it03:82/scripts/wgate/bbpstart/!
or URL:
http://it04:82/scripts/wgate/bbpstart/!
Production cluster URL - http://clusterits20:82/scripts/wgate/bbpstart/!
The problem is as follows :-
'The basic issue is that the users have to click on the 'enter' button on the BBP login screen (for server it04) main BBP screen, (hitting/pressing 'Enter' button doesnt work).
But on server IT03 users can just press enter on the login screen on their keyboards for IT03 (server it03) to login ! ..... so depending on which server they connected to within the cluster address they would have to press or click buttons etc to get past the BBP login screen.'
The templates for bbpstart and bbpglobal for the are identical on both physical servers.
The services on both the servers are also identical.
Please advice on how to solve this issue.
Thanks and regards
Avinash
Edited by: Avinash Rajan on Aug 11, 2008 5:33 PM

Hello Avinash,
Did you check the template bbpglobal/99/login.html? That is the actual file being used for the bbpstart service logon screen. The login.html should contain a function for login using the enter (function loginOnEnter(theEvent)). Maybe this is missing from your it04 server?
I would recommend publishing the bbpglobal Internet Service to your it04 server again.
Edgar

WebLogic 10.3.4 Adminserver Performance Issues - Clustered Environment.

I am running WL 10.3.4 64 bit version on 64 bit hardware (blade serves) running Solaris 10 and I have the servers setup in a cluster. There are 3 machines are part of the cluster.
Managed Server1 is on Blade Server 1
Managed Server2 is on Blade Server 2
Managed Server3 is on Blade Server 3
We have about 20 different clusters setup and 60 servers across those 20 clusters.
The adminserver runs on a separate box (not a blade server). Tasks that do NOT involve the servers or deployments tabs are relatively fast however, when I go to the servers or deployments tab, and to some extent the clusters tab, the response is very slow.
For example, I am running WL 10.3.0 on less robust hardware. I installed the patch that Oracle published to resolve this particular issue. We have a similar number of clusters and servers on in that environment (3 severs to a cluster). The servers tab responds in 10 seconds.
In WL 10.3.4, the servers and deployments tabs take about 45 - 60 seconds to respond. This adds a LOT of overhead to doing any server or deployment related tasks.
I realize that the admin server has to query all the managed servers but performance in 10.3.4 is much worse than WL 10.3.0 with the performance patch.
Has anyone else run into similar performance issues with the WL 10.3.4 ADMIN Console running in a clustered environment? If so, does anyone have any suggestions to help improve performance.
I also run a WL 10.3.4 instance in a single server environment and the servers and deployments tabs respond withing 10 seconds (very fast). It is just in a clustered environment.
Regards,
Mark

hi Mark, it's a very interesting story, any follow-up? I would be curious to know the reason for the delay...

Clustered Servers and Applet (Very URGENT Please)

Hi,
I'm using win2k/CF application server 4.5/IPlanet web server/NN4.79/JRE1.4.1/.
I have an applet in one of my web page which downloads a file from the web server(stored in http://www.abc.net/doc/zip directory on the server).
Now when I run this code on the development server, everything works fine.
But in production environment, we have 3 clustered web servers controlled by a controller (Local Cisco Director) machine. The applet code looks like this:-
strURL = "http://www.abc.net/doc/Zip/"; //the URL from which the zip is to be downloaded. Basically it hits the Controller first
     strZipFileName=getParameter("zipFileName");
     strURL = strURL+strZipFileName;
          //Get Connection to Server and Download the file
          try{
          objUrl = new URL(strURL);
          httpCon = (HttpURLConnection) objUrl.openConnection();
               //DownLoad the File
               InputStream fileInputStrm = httpCon.getInputStream();
               // create a file object which will contain the directory structure to copy in the local machine
               localZipFile = new File(strLocalFilePath+strZipFileName);
               FileOutputStream fileOutStrm = new FileOutputStream(localZipFile);
               int ch;
          while ((ch = fileInputStrm.read()) != -1) {
                    fileOutStrm.write(ch);
               // Close the InputStream, HTTP connection , File stream
          fileInputStrm.close();
          httpCon.disconnect();
          fileOutStrm.close();
Now when the home page of the site is loaded it makes the connection and controller redirects this request to one of the web servers (say 1) which is least loaded.
Now when the applet is loaded, the new connection is established to the controller and it may redirect the request to any one of the 3 web servers (not necessarily to 1 which is required). Now sometimes it hits the right web server on which the required zip file is stored, and sometimes it doesn't hit the correct server.
Any clue what we can do in this scenario so that the applet downloads the correct file from the right server?
Please help it is very urgent....
Thanks

Did you ever get a response to this or figure out how to run in a clustered environment? I am now running into the same issue and would be interested in whatever you learned.

ColdFusion 9 - Verity Issue

Hello. We recently upgraded to CF 9 Enterprise running on Windows Server 2008 64 bit. It is running well with an issue we cannot seem to trace down.
The server works great for a random period of time and then randomly we get this error:
"The collection you specified does not exists or is not registered with the ColdFusion Search Service."
If you login to the CF Admin you see the collection but for some reason it cannot be searched... or accessed. If you attempt to delete it, the admin throws an error stating the collection doesn't exist. What ends up happening is that we recreate it, and repopulate the index. It runs fine for a while and then happens all over again...
Our current work around was to use the old server these sites were operating off of and use a remote web service. It's not ideal but given that the old system (CF 8 Standard on Windows Server 2003 32bit) doesn't have the above issue it "patches" the problem while we investigate a fix.
Any input/help would be appreciated.
Thanks
-Thomas

I like to add the following... when I check the log file verity-install.out I see the follwoing............
/coldfusion/verity/k2/_ilnx21/bin/vconfig: /lib/ld-linux.so.2: bad ELF interpreter: No such file or directory
ERROR: Configuration failed to complete successfuly at Fri Sep 14 08:13:23 PDT 2012

Can a single JNDI tree be shared by multiple, non-clustered servers?

We have a situation involving multiple, but non-clustered, WebLogic servers,
say Server 1 and Server 2. A client needs to access an EJB which happens to
be located only on Server 2. We would prefer that the client not need to
know the exact location of the EJB, but use the naming service hosted by a
designated server (Server 1) to locate the EJB.
At a minimum, this would require that both servers bind their respective
EJB's into a common, shared JNDI naming tree hosted by Server 1.
Is this possible?
John Armstrong
WebLink Wireless, Inc.
[email protected]

hi,
I have not tried this, althoug with the webogic rmi and weblogic jndi
implementation I would guess this is straightforward.
Binding remote objects in WL JNDI binds the stubs only, not the object
itself. The worst case scenario here would be for you to implement proxies
and bind them in JNDI, although I would guess that binding the home object
should give you the required functionality.
Just try it, rebind the home object on a different server than the one the
bean is deployed on and test. My guess is that it will work.
(I repeat : I have not tested this my self, this is just a suggestion)
Regards,
Anders M.
John N. Armstrong <[email protected]> skrev i
meldingsnyheter:3a93faba$[email protected]..
We have a situation involving multiple, but non-clustered, WebLogicservers,
say Server 1 and Server 2. A client needs to access an EJB which happensto
be located only on Server 2. We would prefer that the client not need to
know the exact location of the EJB, but use the naming service hosted by a
designated server (Server 1) to locate the EJB.
At a minimum, this would require that both servers bind their respective
EJB's into a common, shared JNDI naming tree hosted by Server 1.
Is this possible?
John Armstrong
WebLink Wireless, Inc.
[email protected]

Verity issue after purging a collection

All,
I have recently discovered an issue with the way verity
provides suggestions when returning search results.
I have created a verity collection using verity spider. i
have registerd this collection in CFAdmin. I have successfully
searched against this collection and received 'suggestions' where
approriate.
So far so good.
However, when i re-run the spider (which purges the
collection then re-builds it) i no longer receive 'suggestions'
This issue is easily reproducable and consistently has
similar behavoir in that following the initial registering of the
collection it works but subsequent rebuilds of the collection
(which obviously don't need re-registering as the collection is
alreay registered) faiuls to provide suggestions
Am I missing something subtle here?

I'll restate this in simpler fashion.
Can you retrieve MX 7 verity search results on a MX 6.1
server?
Can this be done via CFHTTP or any other similar method?
Thanks

Please help in db connectivity issue after servers moves to diff location

Our database is 10.2.0.3 with 2 nodes in RAC. Our database servers are MS 2003 R2. Recently we moved our database to different servers in different location. Say we have location A, B and C.
A - the location for database users include me.
B - the old db location
C - the new db location.
Location B and C are in the same city but different places. A and B are located in different cities. I was told that there is no direct connection between A and C and the connection between A and C has to go through B. Since the move we have experienced database connectivity issue between A and C. Some users got “ORA-03113: end-of-file on communication channel”. For me I have constantly got “ORA-03135: connection lost contact” from Sqlplus or Toad if I leave my session idle for one hour or longer no matter if it uses dedicated connection or shared connection. We also asked our network admin to make sure there is no firewall in between and we also add KeepAliveTime and KeepAliveInterval to the new db server registry. We also have a TAR with Oracle.
Thanks a lot for your help in advance.
Shirley

It feels like a firewall issue, yes. But this is realistically just a hunch-- your network admins ought to have a lot more tools at their disposal to diagnose the problem. It's quite possible that it's not a firewall but a flaky router or excessive packet loss over a WAN or any number of potential network related hiccups. If I had to place a bet, I'd be wagering on a firewall, but I certainly wouldn't be staking my life on it. I'd definitely be leaning on the network folks to diagnose the issue(s)-- there may well be multiple problems, i.e. a firewall that times out idle connections after an hour and a flaky router that drops connections for some subset of users between A and B.
Justin

Verity Issue - Clustered Servers

Similar Messages

Maybe you are looking for