on 2012 Apr 18 10:29 AM
I am trying to setup a Relay Server in our DMZ so that MobiLink clients can connect to the server from remote locations without the use of a VPN. We do not have a MobiLink or Relay Server Farm and am struggling with what the parameters should look like in the rs.config file.
[backend_farm] Would we include this section in the file?
[backend_server] enable = yes farm=??? id=??? description = MobiLink Server mac = Use the MAC address of the machine runnign the MobiLink server?
Thanks for any help and advice.
Request clarification before answering.
Even though you don't really have a farm, you still need to define the farm with a single back end server. It might look something like :
[options] start=no verbosity=4 shared_mem=40M [relay_server] host=rdomarat-pc2004 enable=yes http_port=80 https_port=443 [backend_farm] id=mlrep.MLFarm enable=yes verbosity=4 [backend_server] id=ml1 farm=mlrep.MLFarm enable=yes verbosity=4 description=ML Server running on rdomarat-pc1001 mac=00-25-64-e9-73-1a
The MAC address you use for the backend server is the MAC Address of the machine where the RSOE is running that connects the ML Server with the Relay Server. This is typically the same machine where the ML Server runs (particularly if you are using the embedded RSOE in ML v12), but it does not have to be.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
The [backend_server] section illustrated above is missing a typical token property but it is indeed optional. In fact, both the token and the mac are optional in the [backend_server] section. Not specifying them means your rsoe command doesn't need a -t and the implication is anyone with an rsoe can pretend to be your backend server, connect to the RS and hijack your clients and stop your legitimate rsoe from connecting if the offending one gets connected first.
Hello. I found similar error in your log file Communication error [SYS1229: An operation was attempted on a nonexistent network connection...] occurred while writing to up channel of backend server 'mp1package' in backend farm 'Afariapackage' . When we reload our server, after reload our server running without errors some time(30-60 minutes, but later we have this problem) Can you help us?
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
We were not able to resolve and have shelved the project for now. Was a little frustrated that this thread died without any additional responses and we could not afford to spend any more time on this effort. We are looking to start working on early in summer if someone from ianywhere can provide any additional insight for us.
Sorry, I'd left on vacation in April after responding to your thread, but nobody jumped in when I left, and I forgot to check back when I returned.
I looked at the log you posted on the 23rd, in context of these errors you were seeing the ML log :
I. 2012-04-19 08:50:21. <main> <oe>: <upchannel-0000> Successfully connected to relay server: MRS0101:80 E. 2012-04-19 08:50:32. <main> [-10354] <oe>: <dnchannel-0000> doHTTPRequest: read reply failed I. 2012-04-19 08:50:38. <main> <oe>: <upchannel-0000> Successfully connected to relay server: MRS0101:80 E. 2012-04-19 08:50:38. <main> [-10354] <oe>: <upchannel-0000> Error: Encountered authentication error with Relay Server, Farm ID or backend server ID could be invalid
In the Relay Server log, we see the RSOE successfully connect, but then there is no further communication from this RSOE, and we soon see another RSOE instance (from the same MAC Address) attempting to reconnect again. Since it's using the same farm_id and server_id of an RSOE that the Relay Server thinks is still connected, it denies the connection.
The ML/RSOE log shows that there was a read error shortly after connecting, but there is no evidence of any error in the Relay Server. This tells me that the IIS Server returned the error before the read request was ever passed to the Relay Server running on IIS. Looking at the IIS Access logs, as well the HTTPERR logs from the time that the RSOE reported the read error would help us figure out why the read failed.
Best guess : I've seen something similar to this when the IIS configuration does not set the maxAllowedContentLength to 2GB in the location section where the Relay Server ISAPI extensions are located.
<location path="Default Web Site/rs"> <system.webserver> <security> <authentication> <anonymousauthentication username=""/> </authentication> <requestfiltering> <requestlimits maxallowedcontentlength="2147483647"/> </requestfiltering> </security> <handlers accesspolicy="Execute, Script"/> </system.webserver> </location>
After the RSOE connects, it makes PUT and GET requests to the IIS server with a 2GB size. If you haven't set the maxAllowedContentLength, then the first liveness packet sent by the RSOE gets a read error and IIS kills the connection.
Here is one of the errors from the log:
E. 2012-04-23 10:49:55. <2344.3588.F0B0Up> RSE3003: Redundant outbound enabler connection for backend server 'MLServer2' in backend farm 'MobiLink.Farm' was ignored
I tried uploading the log file. Hopefully I did it correct:
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
61 | |
8 | |
7 | |
6 | |
6 | |
5 | |
4 | |
4 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.