Could not access server

we are trying to have a .net web services accessing the livelink web services thru ssl via a load balancer. We have configured the certificates and the web.config to accept https. From the server running the webservice we can access thru IE the livelink web service: http://hostname/les-servicesssl/authentication.svc we no certicates error since we have import the CA certificate in the local machine store.

when we start our application and the apps webservices try to open the LL web services, we have an error: "Could not access server" I assume we only need port 443 open on the firewall? We don't need port 2099 open?

Find more posts tagged with

Comments

Appu_Nair1

In the server's web.config I assume you have a right server name and right lapi port?

You may want to look by starting to put wantlapi=true

While not entirely relevant you may want to rule out IPv6 vs IpV4 problems etc to even see

some good quality logs

I got lucky sometime back and noted that in this thread

https://knowledge.opentext.com/knowledge/cs.dll?func=ll&objId=25609521&objAction=viewincontainer

Guy_Pomerleau

We did further test and it is not related with the load balancer because we try to connect directly to a specific server. The issue is kind of random and seem to be related to the token we generate. Since we are in a load balance environmnet we authenticate on each call we do. Sometime we get : "could not access server" what is the lifetime of a token on the livelink server. Is it the LL service or the web service that generate the token? attache is a web service log file. You can notice a pattern: 1 authentication then 2 Documentmanagement.

Anyone with idea. We have the same code internally but we use windows authentication and we don't have that issue, only with livelink account in SSL

Appu_Nair1

In your livelink server if you go to the function - Configure Security Parameters

?func=admin.securityvars do you have each livelink server use the IP address of the server to generate the cookie as in 255:255:255:255

If so each loadbalanced livelink server is sending to your webapp a cookie(token) that can be used only with that.

On a test sytem change that to "Do not Use IP address" and see if that works consistently,but that is very insecure

In a newer livelink Update OT has enabled something called 'x-forwardedfor' headers that is supposedly a way to identify friendly LB farms

and basically work this problem out It is used to load balance another product from ot based on webservices called (RCS)

I have run into this again before hence the nudge

Jason_Smith

The error "Could not access server" means that the socket connection from the application server to Content Server failed. It has nothing to do with the token being used. Unfortunately, the logs will only provide the error message and no further details.

If you are connecting to the correct host using the correct port (which looking at the logs it seems fine to me) and Content Server is running fine (is it possible it was down at all?), I'm not sure why this would be failing.

Kyle_Swidrovich

Hi Guy,

The Content Server service is generating the token.

It’s equivalent to Content Server handing out a “cookie”.

And it’s only the Content Server service that attempts to validate the token.

I don’t have a code-level knowledge of the inner workings of the web service, so I’m not sure what the com.opentext.api.LLIOException “Could not access server” specifically means.

If you don’t see a request coming into a Content Server instance at that time, then it seems like socket connection issues (e.g., IP restrictions, ephemeral port exhaustion on the client, etc).

If you do see a request coming in, then it could mean an issue with the auth token.

In that case, how long is your cookie expiry?

Are the times on each of your servers in sync?

Maybe the cookie/token from Server A is expiring as per Server B’s view of the cookie’s age.

Regards,

Kyle

Guy_Pomerleau

We have also users that have "Could not connect to server" in the livelink web interface more and more often. It started sometime around 18 oct. If I look in the IIS log, we have error 500. see attachement at the end.

Guy_Pomerleau

where and how the cookie expiration work in a context of multi-tier apps, in a load balance livelink environnment:

Web apps ---> web services ---- ssl auth ----> load balancer (sticky session enable) ----> livelink web services (2 instance per server on 2 different ip and 2 servers so 4 instance total)