[Spce-user] Observations/issues after upgrading to 3.6.1 - nginx

[Alex Lutay]

Dear Thomas,

I suppose we have found the reason of your problems.
ngcp-update-cfg-schema and ngcp-update-db-schema were not executed on 
your system because /etc/ngcp_nodename is wrong.

CE edition must have "spce" in /etc/ngcp_nodename as well as into DB 
(ngcp.*_schema.node)

Initially we used hostname as "node" value (ngcp.*_schema.node) which
created a lot of problems as hostnames were changed time-2-time.

So, we created /etc/ngcp_nodename which was automatically initialized as 
"spce" for CE (I suppose you have manually changes it to "sip0"):
> upgrade-mr3.3.*-to-mr3.4.1:  echo "spce" > /etc/ngcp_nodename

Also we were shared several times the proper way to fix old systems:
https://lists.sipwise.com/pipermail/spce-user/2014-September/007257.html

The „active node“ is being used in PRO product where you have two nodes
controlled by Heartbeat and provided high-availability for you.
In this case DB/cfg updates must be executed on inactive node only.

ngcp-update-db/cfg-schema checks CE by "spce" and ignores CE activity:
> running_on_active_node() {
>   if [ "$nodename" = "spce" ] ; then
>     return 1
>   else

So, I would recommend you the following:
backup your current configs (just in case)
restore config from upgrade backup
run the following commands:
> mysql -u root ngcp -e 'update db_schema set node = "spce"; update cfg_schema set node = "spce";'
> echo "spce" > /etc/ngcp_nodename
> ngcp-update-cfg-schema
> ngcp-update-db-schema
> ngcpcfg apply

After that your system should be in proper state.

---

Also I do recommend you to check:
> #hostname -f
> iparlo.net

"-f" must return FQDN (Fully Qualified Domain Name), so probably it
should be "sip0.iparlo.net" for your case.

On 17/11/14 16:36, Thomas Odorfer wrote:
> Some checks:
>
> # ngcp-check_active -p
> !
> # ngcp-nodename
> sip0
> (identical with ngcp.cfg_schema.node )

-- 
Alexander Lutay