Several resets until GD gets service.

[enio.eltz]

Hi

I'm having a problem with a GD. It resets several times before getting service.

The environment is:

OXE Aplliance Server R7.1-f5.401-33-br-c0s1.
There are 8 MG ( 1 of them for voice mail ).

The problem happens with shelves 14 and 15, 14 is the MG and 15 its expansion. Everything works fine when GD is in service. When GD loses signaling with AS, it resets. When GD gets signaling from AS again, it will get in service, but several times I have the error below:

loading MCV module - please wait ...
DSP firmware successfully read in flash (size=754156)


/mnt/flash/local/startemg: OK

Startup completed - awaiting user login ...


Rack 0 is a large

Starting /etc/rcS.d/99crond

Welcome to e-MGD
IP is 10.0.8.200
eMGD login: Start VG_PROCESS using /dev/rtf5 and /dev/rtf4
Trace OFF
Inserting country module : country_param_BR.o
country_param_BR.o.tar.gz
country_param_BR.o
rmmod eventuel avant insmod : country_param_BR
rmmod: module country_param_BR is not loaded
puis appel de insmod : country_param_BR.o


====== record_country : 35 (BRAZIL)


BCP_Resource::create_flow_control_handler --- DSP number = 0

MGD - RACK CONSTRUCTION


Rack 1 is a large

SCS CNX is nok : e-MGD Reset



Reset: startemg downloading from SCS failed: e-MGD Reset (0 - 0)

I have done a test turning EXP1 off, in this case it doesn't reset, shelf 14 starts ok. With the shelves 14 - MG and 15 - EXP1, it resets several times until getting service. After getting service everything works fine.
Does anybody know this error "Reset: startemg downloading from SCS failed: e-MGD Reset"?
INCVISU doesn't show an incidents, only it loses shelf 14.

Thanks.

[smileyman]

Hi enio.eltz,

there was the same problem. - But without solution
Is the IP-Configuration right? (Gateway, Netzmask, etc.)
Why does the shelf reboot? - Exist the same problem, if you reset the set via rstcpl or only on a IP-link-problem?

Best regards
smileyman

[knightrider]

have you tried to manually recharge the GD with correct binaries?

[enio.eltz]

Hi Knightrider

I have already tried. It happens the same thing. I have tried with another GD too.

Hi Smileyman

I have read that post. In this case I don't have 9 series there. I just have IPTouch.

I did some tests in lab this afternoon. I have in lab an OXE R9.1 on VMware with one GD an one expansion rack. I discovered the following:
When I do the command shutdown on CS, GD shows the same message SCS CNX is nok : e-MGD Reset
Reset: startemg downloading from SCS failed: e-MGD Reset (0 - 0).
I understood SCS CNX is nok like Standard Call Server Connex is not ok, of course, I turned it off. And GD tries to start download of startemg again but it can't because CS is turned off.
At my customer I can have a problem with the link MPLS or something like this.
When GD is starting, it is downloading files from CS to start the boards, in a moment it can't receive an answer of keep alive from CS for example, and it resets again.
As I said before, when I turn the expansion off I don't have the problem. MG has 2 SLI16, 1 T2, 1 APA8. Expansion has 4 SLI16 and 1 UAI16 with 4 ERB IBS.
I need to do more tests at the customer. I post here news.

Thanks

[Mr T]

Sounds like an issue with the IP network.

[enio.eltz]

Hi Mr T

I agree with you. I am gonna do more tests at the customer.

Regards.

[Konstantinos.E]

helloz

go to ip-int/ip parameters-Ethernet address checked by tftp make it false...and see what happents..

[enio.eltz]

Hi folks

I am resuming work in this problem. Last saturday I did some tests at the customers where I could find something that can help me to solve the problem.
As I told before, when I turned expansion rack off, that MG starts rightly.

I did:

1 - I removed all SLI16 boards from the expansion rack (MG 15) and kept one UAI16. MG 14 (main) and 15 (expansion) started OK.
2 - I connected the first SLI16 and did rstcpl 14 0. MG 14 (main) and 15 (expansion) started OK.
3 - I connected the second SLI16 and did rstcpl 14 0. MG 14 (main) and 15 (expansion) started OK.
4 - I connected the third SLI16 and did rstcpl 14 0. MG 14 (main) and 15 (expansion) started OK.
5 - I connected the fourth SLI16 and did rstcpl 14 0. MG 14 (main) and 15 (expansion) didn't start OK. GD config port showed this:

loading MCV module - please wait ...
DSP firmware successfully read in flash (size=754156)
/mnt/flash/local/startemg: OK
Startup completed - awaiting user login ...
Rack 0 is a large
Starting /etc/rcS.d/99crond
Welcome to e-MGD
IP is 10.0.8.200
eMGD login: Start VG_PROCESS using /dev/rtf5 and /dev/rtf4
Trace OFF
Inserting country module : country_param_BR.o
country_param_BR.o.tar.gz
country_param_BR.o
rmmod eventuel avant insmod : country_param_BR
rmmod: module country_param_BR is not loaded
puis appel de insmod : country_param_BR.o
====== record_country : 35 (BRAZIL)
BCP_Resource::create_flow_control_handler --- DSP number = 0
MGD - RACK CONSTRUCTION
Rack 1 is a large
SCS CNX is nok : e-MGD Reset
Reset: startemg downloading from SCS failed: e-MGD Reset (0 - 0)

6 - I tried to change one SLI16 from MG 15 (expansion) to MG 14 (main) and did rstcpl 14 0. MG 14 (main) and 15 (expansion) didn't start OK.
7 - I removed one SLI16 from the cabinet. MG 14 (main) and 15 (expansion) started OK.

In other words, MG 14 an 15 just start when I keep plugged 5 SLI16 (MG14 + MG15), when I have 6 SLI16 plugged MGs don't start rightly.
I don't rebember but it doesn't have limit of 80 analog sets with two rack L. If it were one rack L, it would be 80.

Does anybody have any opinion??

Thanks.