Project

General

Profile

Feature #2618

write Osmocom Network In The Box wiki page

Added by neels 17 days ago. Updated 1 day ago.

Status:
Resolved
Priority:
Urgent
Assignee:
Target version:
-
Start date:
11/07/2017
Due date:
% Done:

100%

Spec Reference:

Description

Describe how to run a NITB with the new split components.


Related issues

Related to Cellular Infrastructure - Feature #2584: Have "osmo-network-check" to verify "Network In The Box" setup New 10/20/2017
Related to Cellular Infrastructure - Support #1719: review different 'getting started' guides in the wiki and unify/update them Stalled 05/17/2016
Related to OsmoHNBGW - Bug #2638: segfault during IuCS connect request Resolved 11/14/2017

History

#1 Updated by neels 17 days ago

Osmocom Network In The Box is written, but it is not completely verified to be correct yet.
Open issues are

  • use of osmo-mgw vs osmo-bsc_mgcp,
  • the GGSN tun device config,
  • verify that those minimal configs are indeed sufficient, and submit as config files to the various git repositories.

#2 Updated by neels 17 days ago

  • % Done changed from 0 to 50

#3 Updated by neels 17 days ago

  • Status changed from New to In Progress

#4 Updated by neels 17 days ago

  • Related to Feature #2584: Have "osmo-network-check" to verify "Network In The Box" setup added

#5 Updated by neels 17 days ago

  • Related to Support #1719: review different 'getting started' guides in the wiki and unify/update them added

#6 Updated by laforge 16 days ago

On Tue, Nov 07, 2017 at 12:50:18PM +0000, neels [REDMINE] wrote:

  • the GGSN tun device config,
  • verify that those minimal configs are indeed sufficient, and submit as config files to the various git repositories.

could that actually be tested automatically? Now with osmo-bts-virtual etc. we could actually
take a set of config files, start all processes and see if a LU succeeeds? Should that be
a separate gerrit job, or should we start adding a osmo-bts-virtual to osmo-gsm-tester? As
there's no ofono for "mobile", probably the latter won't work yet until we have the primitive
+ LUA scripting in there :/

#7 Updated by laforge 16 days ago

  • Priority changed from Normal to Urgent

#8 Updated by neels 16 days ago

laforge wrote:

  • verify that those minimal configs are indeed sufficient, and submit as config files to the various git repositories.

could that actually be tested automatically?

sure. If it runs on jenkins, it should be in a docker, so that we can freely use network interfaces; which again touches on various other topics: should we "duplicate" the osmo-gsm-tester setup? should we move osmo-gsm-tester to use docker images? should be move all jenkins build slaves to docker images?

My first step is still to go through all the components and find the minimal working config.

I'm taking slight detours to polish up some rough edges I'm finding there, and verifying the config is taking way longer than I anticipated. After all this is the user's perspective on everything, and I'm reflecting on pretty much all settings, parsing and documentation... I didn't realise that at first.

should we start adding a osmo-bts-virtual to osmo-gsm-tester?

Sounds like that makes sense, adding another BTS model to it to benefit from the remaining setup already present.
The test scripts can also use non-ofono ways to trigger virtual modem side actions, no restraints there really.
If ofono can interface with it though, we can presumably re-use the same test cases for the virtual BTS.

I am not familiar with:

there's no ofono for "mobile", probably the latter won't work yet until we have the primitive
+ LUA scripting in there :/

I guess after your presentation on osmo-bts-virtual next week I'll have a better grasp on it?

#9 Updated by neels 4 days ago

  • % Done changed from 50 to 60

I've verified 2G voice and data operation as described on the wiki page, but I am still thoroughly stumped on osmo-hnbgw not working for me at the moment. osmo-hnbgw crashes as soon as a subscriber attempts to attach to the network via 3G, see #2638

#10 Updated by neels 4 days ago

  • Related to Bug #2638: segfault during IuCS connect request added

#11 Updated by neels 1 day ago

  • Status changed from In Progress to Resolved
  • % Done changed from 60 to 100

build problems figured out, segfaults resolved, config verified.

Also available in: Atom PDF