Hi, List!
I'm trying to connect OpenBSC with a real MSC. Unfortunately I have many problems. The first is, that every minute or so, I got these errors:
<000f> bsc_msc_ip.c:1047 The connection to the MSC was lost. <000f> bsc_msc_ip.c:976 Lost MSC connection. Freing stuff. <000f> bsc_msc.c:236 Attempting to reconnect to the MSC. <000f> bsc_msc.c:230 Attempting to reconnect to the MSC. <000f> bsc_msc.c:137 Attempting to connect MSC at 192.168.2.103:5000 <000f> bsc_msc.c:170 MSC Connection in progress <000f> bsc_msc.c:99 (Re)Connected to the MSC. <000f> bssap.c:92 Reset ACK from MSC
I think, it is not normal, that the communication with the MSC was lost. Both of them (OpenBSC and MSC) are on the same Gigabit network, so there is no connection problem... Any idea?
Another problem: I get very oft the error:
<0000> chan_alloc.c:441 (bts=1,trx=0,ts=0,ss=0) Recycling Channel.
Any idea?
Thanks a lot!
On Thu, Dec 16, 2010 at 01:15:12PM +0100, Luca Bertoncello wrote:
I'm trying to connect OpenBSC with a real MSC. Unfortunately I have many problems.
this is to be expected. OpenBSC (osmo-bsc) has only been tested with one particular version of one particular MSC vendor so far. The only version that we ever had access to.
So I think getting it to work with a new/different type of MSC will most likely need some kind of adoption. Maybe not as much work as adding support for a new BTS vendor/model, but definitely some work.
Also, from our experience with that one MSC: Zecke had to work around quite a number of bugs in the MSC. Whether the same bugs or workarounds work with your MSC is of course unknown.
Any idea?
No, I think you will have to debug this yourself by looking at the A interface prtoocol traces in wireshark.
On 12/16/2010 04:38 PM, Harald Welte wrote:
On Thu, Dec 16, 2010 at 01:15:12PM +0100, Luca Bertoncello wrote:
Also, from our experience with that one MSC: Zecke had to work around quite a number of bugs in the MSC. Whether the same bugs or workarounds work with your MSC is of course unknown.
Not that many, some more due my own stupidity, but it is true that the code is written to work with this one MSC but I have not really seen a spec violation (it can do some stuff that it should not do though).
On 12/16/2010 01:15 PM, Luca Bertoncello wrote:
Hi, List!
I think, it is not normal, that the communication with the MSC was lost. Both of them (OpenBSC and MSC) are on the same Gigabit network, so there is no connection problem... Any idea?
Well, who is closing the TCP connection? Is it the MSC that is unhappy? Is it the BSC that is unhappy (e.g. you removed a log line talking about timeouts?)? Do you see any other messages in wireshark? Which MSC is that>
Another problem: I get very oft the error:
<0000> chan_alloc.c:441 (bts=1,trx=0,ts=0,ss=0) Recycling Channel.
Any idea?
Why do you think it is an error? It is a normal notice on channel release. Actually you must be using on-waves/bsc-master as well.
-
Harald Welte -
Holger Hans Peter Freyther -
Luca Bertoncello