On Fri, Jul 12, 2019 at 10:26:04AM -0300, Rafael Diniz wrote: > Today I'm working on updating my NuRAN unit with the new osmo stack Can you clarify which particular model/unit that is? I would assume that the 'nightly' OE packages sysmocom provides should have all related features.

> I'm writing just in case someone already put online a set of config > files with the parameters needed by such behavior set? > ; ) I would be careful with full configuration files, as they contain all kinds of settings which may or may not do what you want. The settings for subcsriber-create-on-demand are only 2-3, AFAIR. > ps: I have the new splip stack working fine here, I just need to modify > my setup with new features to support subscriber_create_on_demand. Then please simply use the config files you have and not start with something else just because you need to modify one or very few lines... See https://osmocom.org/issues/2542 where Oliver actually also points to a short new chapter in the manual at https://ftp.osmocom.org/docs/latest/osmohlr-usermanual.pdf If yo have more specific questions, feel free to raise them here :)

"check-imei-rqd early", not "check-imei early" On 7/15/19 10:05 AM, Oliver Smith wrote: > Hey Rafael, > > On 7/13/19 5:55 PM, Rafael Diniz wrote: >> Hi Harald, >> >>> On Fri, Jul 12, 2019 at 10:26:04AM -0300, Rafael Diniz wrote: >>>> Today I'm working on updating my NuRAN unit with the new osmo stack >>> >>> Can you clarify which particular model/unit that is? >>> >>> I would assume that the 'nightly' OE packages sysmocom provides should have >>> all related features. >> >> It's a Nuran LC 1.5. I'm using debian 9 armhf packages and compiling by >> hand the BTS and PCU against LC 1.5 headers. >> >>>> I'm writing just in case someone already put online a set of config >>>> files with the parameters needed by such behavior set? >>>> ; ) >>> >>> I would be careful with full configuration files, as they contain all kinds >>> of settings which may or may not do what you want. The settings for >>> subcsriber-create-on-demand are only 2-3, AFAIR. >>> >>>> ps: I have the new splip stack working fine here, I just need to modify >>>> my setup with new features to support subscriber_create_on_demand. >>> >>> Then please simply use the config files you have and not start with >>> something else just because you need to modify one or very few lines... >>> >>> See https://osmocom.org/issues/2542 where Oliver actually also points to >>> a short new chapter in the manual at https://ftp.osmocom.org/docs/latest/osmohlr-usermanual.pdf >>> >>> If yo have more specific questions, feel free to raise them here :) >> >> Fine, I'll just adapt my config - Thanks!! > > As Harald pointed out, the documentation could use some more examples. > I'll update the docs accordingly, but to make your life easier, here are > the configuration options relevant for your use case (with sending the > IMEI to the HLR). > > osmo-msc.cfg: > msc > check-imei early > > osmo-hlr.cfg: > hlr > subscriber-create-on-demand 5 none > store-imei > > > This will create 5 digit MSISDNs for the subscribers, and disable CS NAM > and PS NAM by default (circuit switched and packet switched network > access modes). Subscribers can't make phone calls or use cellular data, > so their phones will not connect to the network, but the entry in the > HLR will be created. > > After a new subscriber was created, it will look like this in the VTY > (I've replaced the real IMEI and IMSI with zeros): > > OsmoHLR> enable > OsmoHLR# subscriber imei 00000000000000 show > ID: 1 > IMSI: 000000000000000 > MSISDN: 58192 <- randomly generated > IMEI: 000000000000000 > CS disabled > PS disabled > > The idea is now to enable CS and PS for the IMEIs that you know. Right > now, the documentation says: > >> In order to do that, one can set the default NAM to none and manually >> approve new subscribers by enabling their nam_cs and nam_ps parameters >> (e.g. over the VTY). > > But as I was about to create an example of how these commands look like, > I realized that VTY commands for changing nam_cs and nam_ps don't > actually exist yet :\ So for now, these flags have to be changed > manually in the sqlite database... change nam_cs and nam_ps to 1 in the > subscriber table. Maybe this is still useful for testing. I will add the > missing VTY commands shortly. > >> >> Cheers, >> Rafael Diniz >> > > Cheers, > Oliver >