we have a deployment with a TX board with two linksets, each with four links.
We have developed a TCAP application for INAP which receives transaction BEGINS and is answering with CONTINUE and END messages. Everything works fine. But we see in the mtpmgr, that our messages are always sent on the first link. The other 3 links of the linkset are nearly never used. We have set tInfo.qos.seqDlvy = 1. The received messages are uniformly arriving over the 4 links.
How can we ensure a smooth load over all the 4 links of the linkset?
You didn't mention your SS7 software version. If you're using SS7 5.3, it's a bug. You should ask you distributor (or Dialogic directly, if you have a support contract) for patch 9413 (description "SS7 SCCP - Fix Class 1 SLS Override").
Best Regards, Greg Blumenthal
thank you for your fast reply.
We're using SS7 in version 5.1.1. Is the patch also required for this version?
There is another patch for SS7 5.1.1: Patch 9163, description "SS7 SCCP - 8-bit SLS change restored". That description doesn't sound too much like your issue, but I see a note with the patch that says "Symptoms of this problem would include TCAP messages being sent on a single link.", so it should be the one you need.
Please note that SS7 5.1 and 5.1.1 are near the end of maintenance (July 20th, 2012). So you should be looking to move to SS7 5.3 relatively soon. Please see http://www.dialogic.com/Products/signaling-and-ss7-components/download/tx-software.aspx
these patches work fine (we've tested both NA versions). Thank you!
Now we can continue our load tests and we have found a new problem. The board seems to hang under high load (mixed ISUP calls / TCAP transactions; it is an INAP assisting IP). When we try to locate the problem, we see, that we can't use sccpmgr.exe . Each command is blocking. So we assume that the SCCP layer is hanging.
Do you have any idea about this reproducable behaviour (we're using NA 5.3 / Windows)?
Best regards, Michael
Thanks for the update on the patches. Glad they solved your load sharing issue.
The new symptom you reported doesn't sound familiar. I think you had better contact Dialogic support. Usually for this kind of thing, the key will be determining the details of how to reproduce it, by understanding the patterns. For example, does it happen after a certain amount of time? After a certain number of transactions? Immediately when the load hits X CPU usage? A txalarm output may also be useful.
But I doubt you'll get a solution to this kind of issue on the forums. It's the kind of thing that will have to be formally reported to Dialogic for further investigation.
Michael Kloos:The board seems to hang under high load (mixed ISUP calls / TCAP transactions; it is an INAP assisting IP).
You could check TX board free memory, cpu and error log with cpcon utility. Commands that you need to execute are - log, util, mem, pools, usage. See TX Utilities Manual
Check also that you have enough ss7 links, depending on the load 8 links may be insufficient.