Anyone had expenerice of TN 4.5.1 and Sequelink 5.1 hanging? …
We are running the B2B 4.0.2 server with TN 4.5.1 on win2000. The B2B server is running properly, but TN was not logging data to the database, and all the service that were connecting to database were hanging(for example routeBizdoc). When I checked the Sequelink service(Oracle) status it was showing ‘green’ signal, but it was not responding when i tried to open the monitoring details. There was no error log in B2B server and Sequelink server. After restarting both the servers it is working now.
I’m tring to figure out the cause and immediate solution for this issue. Any information from you would be appreciated.
We were also facing a similar problem.
Below is an excerpt from a communication with webMethods support…
But we are using 4.6 IS/TN and 5.2 SL.
Sindhu,
As you note, re-starting sequelink server is often required when there was a ‘traumatic’ interruption in its connection (associated with your stopping and restarting Integration Server) might have occured.
I have also noted that at times, WM server doesn’t start at all.
When we try to restart the server after a crash, on a couple of occasions, the server woudn’t start at all.
At this point, when i see the server log. The server.log is stuck at the following sentence and does go no further.
00002E [B2BUNKW.TRANET.1318] Checking data model version information [TRNSERV.000016.000070]
This problem is not resolved until i go and restart the sequelink service(installed as a windows NT service) on the database server.
Note that at this point the oracle server is up and running.
Now again the sequelink service would not respond to to a restart and a machine reboot is required to get the system up and running.
Pls let me know what could we do to resolve this problem.
Not sure if this might be some of the problem, but if you don’t update the connection pool from the out of the box default of 10 you can “lock up” SequeLink at around 30 connected users [Depending on what the services are doing]. To change this add:
tn.store.pool.max=30
If the packages\WmTN\config\properties.cnf file.
Doing so stabalized our systems and we have pushed it to over 70 simultaneous users [TN incoming documents] without a problem.
I’ve seen Sequellink hang thrice in about two years. When this happens, there is no option except to restart (i.e. reboot) the Win2K box on which SL is running. I think this is a bug with SL itself. We use SL 5.1 or 5.2 I think.
Also, this message from WM Support is not true IMHO:
> As you note, re-starting sequelink server is often required > when there was a ‘traumatic’ interruption in its connection > (associated with your stopping and restarting Integration > Server) might have occured.
>
> This is a normal characterisitc of sequelink.
I’ve restarted various versions of IS servers hundreds of times, and never needed a SL restart because of that.
I’d like to qualify my comment about Sequellink never hanging: our Sequellink server proxies various concurrently-running IS servers - test, dev, prod, etc. So, when one IS/TN server is down, the others are still running. Hence, what the WM Support person said may be true for the case of an SL server proxying a single IS/TN server.