You are on page 1of 39

Editor’s Note Maintenance Experience

Quarterly for IMS Core Network Products


No. 3 Issue 271, May 2012

Adhering to the “sincere service for customers” customer


service philosophy, making ZTE products create more value for Maintenance Experience
customers, providing better service for customers, providing the Editorial Committee
best and most practical maintenance experience and methods
for customers, improving service knowledge and skills of related
Director: Qiu Weizhao
professional and technical personnel, and here we come to fulfill
Deputy Director: Zeng Li
the above missions with this volume of Maintenance Experience, Editors:
which is for ZTE IMS core network products. Fang Xi, Wang Zhaozheng, Xu Xinyong,
With the growing fierce competition of telecom market, high Zhang Jian, Zhang Jiebin, Zhao Cen,
quality telecom network is the basic of customer service, and Zhou Guifeng, Xiao Shuqing, Ge Jun,
every network operator pays important attention to equipment’s Zhao Haitao, Huang Ying, Xu Zhijun,
maintenance. This volume of Maintenance Experience contains 18 Jiang Haijun, Dong Yemin, Dong Wenbin
articles of ZTE IMS core network maintenance experience; most
of these articles come from the outfield application failure of ZTE Technical Senior Editors:
IMS products, which should have a significant value for tackling Li Peng, Wang Fangfang, Zhu Xiaopei
with IMS related problems. We sincerely hope this journal will help
you a lot for your maintenance work. Executive Editor:
If you would like to subscribe to this magazine (electronic Chen Xinglong
version) or review additional articles and relevant technical
materials concerning ZTE products, please visit the technical
Maintenance Experience
support website of ZTE CORPORATION (http://ensupport.zte.
Newsroom
com.cn).
If you have any ideas and suggestions or want to offer your
Address: ZTE Plaza,No. 55, Hi-tech Road
contributions, you can contact us at any time via the following
South, ShenZhen, P.R.China
email: doc@zte.com.cn.
Postal code: 518057
Thank you for making ZTE a part of your telecom experience!
Contact: Ning Jiating
Tel: +86-755-26776049
Maintenance Experience Editorial Committee Fax: +86-755-26772236
ZTE Corporation Document support Email: doc@zte.com.cn
May, 2012
Technical support website: http://ensupport.
zte.com.cn
Contents

Technical Specials
IMS Off-line Billing Overview........................................................................................................................2
Introduction to IMS NMS Security Management..........................................................................................3
IMS On-line and Registered Users Query....................................................................................................5

Maintenance Instances
Undefined Operation is Prompted After an SSS Command is Executed.....................................................8
Failed to Log in Because the Default USCC Password is Modified.............................................................10
IMS Processes 200 Telephone Problems.....................................................................................................10
SCSCF Fails to Route Numbers in user=phone Format..............................................................................14
Unregistered Callee Gave No Reply............................................................................................................16
Registration Flow Terminal Fails to Subscribe to UP10................................................................................20
Failed to Call User=phone Number..............................................................................................................21
TD Mobile Phone Fails to Dial IMS Video....................................................................................................23
503 Error is Return.......................................................................................................................................27
LINUX Establishes a Mounting Partition Larger Than 2T.............................................................................30
SIP Error Code 422......................................................................................................................................31
The PP Board Fails to Load Version Properly..............................................................................................33
SSS Modifies Play-Tone by Interior Failure Code........................................................................................35
May 2012 Issue 271 Technical Specials

IMS Off-line Billing Overview


⊙Sun Ming / ZTE Corporation

1 Overview 1. The Signal Main Processor (SMP) of the IMS


The IMS billing mechanism falls into NE monitors network resource occupation status,
off-line billing and on-line billing. and generates AC counting Requests (ACR).
● In the case of off-line billing, 2. The IMS NE sends ACRs to the ZXUN CG.
billing information does not affect related 3. After receiving ACRs, the ZXUN CG replies
services in real time. The ZXUN CG is a Accounting Answer (ACA) to the IMS NE, indicating
ZTE IMS off-line billing device. that the ACR message is received.
● In the case of on-line charging, 4. The ZXUN CG saves received ACRs into the
billing information affects related services original ACR file.
in real time. As a result, the billing system 5. The ZXUN CG filters , screens , sorts and
must interact with the session/service constructs CDRs.
control system. The OCS is an IMS on-line 6. The ZXUN CG removes repeated ACRs,
billing device. constructs corresponding CDRs, and saves them
in the original CDR file.
2 ZXUN CG Functions 7. After combining, associating, filtering some
IMS CDRs fail to satisfy the CDRs, the ZXUN CG generates complete CDRs
requirements of the billing center in and saves CDRs in the final CDR file.
content and format. Therefore, CDRs must
be pre-processed before transmitted to the 4 ACR Overview
billing center. Between the IMS NE and For a session-irrelavant event, such as a
the billing center, the ZXUN CG buffers message, the IMS NE only sends an ACR. After
and pre-processes CDRs. The ZXUN CG receiving this ACR, the ZXUN CG generates a
shields the differences among IMS NEs CDR and save the CDR into the CDR file. For a
from different vendors for the billing center, session-relavant , such as a call, the IMS NE sends
and eases the processing burden on the several ACRs (Start/Interim/ Stop ACR).
billing center. The billing center works out 1. Start ACR indicates the start of a session.
the final billing bills based on the expense 2. If billing condition varies during a session,
settings of carriers. such as media change or the billing interim timer
The ZXUN CG can process at most expires, the IMS NE will send an Interim ACR.
15000 CDRs per second, and collects, 3. When a session is completed, the IMS NE
constructs, combine, sorts, filters, stores, sends a Stop ACR to end the session,
and transmits CDRs. In conclusion, for a session, an IMS NE may
send a Start ACR, no or multiple Interim ACR, and
3 CDR Processing Flow a Stop ACR.
The following flow shows how the After receiving all related ACRs of a session,
ZXUN CG collects IMS CDRs through the the ZXUN CG extracts corresponding billing
Rf interface: information, and finally generates CDRs. ■

 Maintenance Experience
www.zte.com.cn

Introduction to IMS NMS Security


Management
Sun Ming / ZTE Corporation

1 Overview ● Command set operation


Security management ensures that only legal ● Password, unlocking and policy
users can log in to the system, and implements operation
user and role management. Security management The server implements logical
prevents illegal users accessing a system, and processing of login authentication and
provides operators with security control through operation authentication.
operation authentication methods.
Security management is implemented by the 2 Interior Relations of
cooperation of the Web client and the server. Security Management
The clients implement data disposal and display User, role and command set together
functions in the login window and the Security construct the affiliation relations of security
Management window, including: management.
● User operation Affiliation relations among user, role
● Role operation and command set are shown in Figure 1.

Figure 1. Affiliation Relations Between User, Role and Command Set

IMS Core Network Products 


May 2012 Issue 271 Technical Specials

3 User Define the operation command set by adding a


Users are a set of operators who finally command set for a role.
log in to the NM system and perform 4.1 Initiated Role
related operations. The following roles are created in the ZXUN
When creating a user, the administrator SSS while the system is initiating, as shown in
determines the permission of the user by Table 1.
defining the role of the user. 4.2 Role Command Set
You can manage users by defining The role command set of the administrator is
users’ passwords, and setting whether to listed in Table 2.
restrict date, time segment and week. The role command set of the operator is listed
in Table 3.
4 Role The role command set of the maintenance
A role implements operation permission personnel is listed in Table 4.
of a group of users by defining an operation The role command set of the supervisor is
command set. listed in Table 5. ■
Table 1. Initiated Role

Role ID Role Name Role Description Effective or not


1 Administrator The supreme permission is administrator’s permission Y
2 Operator The supreme permission is data configuration permission Y
Maintenance
3 The supreme permission is maintenance permission Y
personnel
4 Supervisor The supreme permission is data view permission Y

Table 2. Role Command Set of the Administrator

Role Name Role Description Role ID Command Set Name

The supreme 1 The command set of 0&M permission


permission is 2 The command set of data configuration
Administrator
administrator’s 3 The command set of system maintenance
permission 4 The command set of data view

Table 3. Role Command Set of the Operator


Role Name Role Description Role ID Command Set Name

The supreme 2 The command set of data configuration


Operator permission is data 3 The command set of system maintenance
configuration 4 The command set of data view

Table 4. Role Command Set of the Maintenance Personnel


Role Name Role Description Role ID Command Set Name
The supreme 3 The command set of system maintenance
Maintenance
permission is system
Personnel 4 The command set of data view
maintenance

Table 5. Role Command Set of the Supervisor


Role Name Role Description Role ID Command Set Name
The supreme
Supervisor permission 4 The command set of data view
is data view

 Maintenance Experience
www.zte.com.cn

IMS On-line and Registered Users Query


⊙Li Peng / ZTE Corporation

You can use different methods to query the


number of the on-line IMS users and the number
of registered IMS users. The followings are several
methods to query the numbers.

1 AGCF NE
(1) Query the number of the users already
allocated with numbers.
Use the probe to query the R_SUB table, the
number of the users is equal to the number of the
users allocated with number on the AGCF, as shown
in Figure 1.
Figure 1. Querying the Number of Users
(2) Query the number of the registered users.
Already Allocated with Numbers on AGCF
Method one: run the query agcf regnum command.
Command result:

Command(No.1):query agcf regnum


Registered Subscriber Number
Unregistered State Subscriber Number
--------------------------------------
0 30
--------------------------------------
1 Record(s)
Execute command successfully(elapsed
0.074s)

Execute command successfully (elapsed


Figure 2. Querying the Number of Registered
0.074s).
Users on AGCF
Method two: use the probe to log in to the front
end and sort the users. Query the R_PUI table to
see the number of on-line PUI users, as shown
in Figure 2. You obtain the number of registered
users by dividing the number by 2.

2 SSS NE
(1) Query the number of the users already
allocated with numbers.
Method one: query the number by performance
statistics after creating the statistics of number of users.

IMS Core Network Products 


May 2012 Issue 271 Technical Specials

a. Create a task for statistics on the number of OSU users, as shown in Figure 3.

Figure 3. Creating a Task for Statistics on the Number of OSU Users

b. Query the existing measurement tasks, as shown in Figure 4.

Figure 4. Querying the Existing Measurement Tasks

c. Query the statistics, and query the number of the users already allocated with numbers, as
shown in Figure 5.

Figure 5. Querying the Number of the Users Already Allocated With Numbers

 Maintenance Experience
www.zte.com.cn

Method Two: Use the command to query the number of the users already allocated with
numbers, as shown in Figure 6.

Figure 6. Querying Number of the Users Already allocated With Numbers

(2) Query the number of on-line users.


Method One: Use the same query method as above. Execute the SHOW SRV PUI
STATISTICS command, as shown in Figure 7.
Method Two: Use the same query method to access the front-end. View the R_PUIINFO table
to query the number of registered users. ■

Figure 7. Query the Number of On-Line Users

IMS Core Network Products 


May 2012 Issue 271 Maintenance Instances

Undefined Operation is Prompted After


an SSS Command is Executed
⊙Miao Feng / ZTE Corporation

1 Fault Symptom by current EMS version, as shown in Figure 1. The


To comply with the group billing current loaded SSS model is ZXUN-SSS(OMM)
091 CDR format specifications, SSS [V4.00.30.B050.0].
patches ZXUN SSS V4.00.30.P7 with the 4. View the model in the /ums-server/works/
SSS V4.00. 30.3.P13.B4.hotpatch010. global/runtime/operationmodel/original/
Nevertheless, after selecting the SSS NE directory, as shown in Figure 3. The latest module
on the U31 maintenance terminal, all the loaded by corresponding SSS version is ZXUN-
configuration commands are unavailable, SSS(OMM) [V4.00.30.3.P13.B40.0]. You must
and the system prompts that the operation reload the version.
is undefined, as shown in Figure 1. 5. Model Loading Flow:
(1) Delete the mml zip file of the wrong
2 Troubleshooting version. For example, if the wrong zip file is
1. After logging in to the SSS ZXUN-SSS(OMM)_V4.00.30.B050.0-configfiles.
maintenance terminal, the engineer can zip, you need to delete ums-server/works/
successfully execute the same commands global/runtime/operationmodel/original/ZXUN-
in the configuration management. That is SSS(OMM)_V4.00.30.B050.0-configfiles.zip.
to say, the SSS NE is not faulty. (2) Delete the cache. Delete /ums-server/
2. On the U31, the engineer can works/global/runtime/operationmodel/server/
successfully excute the configuration uca-mml-model.desc and /ums-server/works/
management commands in the P-CSCF NE. global/runtime/ppus/uca.ppu/uca-role.pmu/
Maybe the SSS model loading is faulty. ZXUN-SSS(OMM)[V4.00.30.B050.0].
3. View the logs in the directory (3) Stop the server and restart the service
export/netnumen/ums-server/utils/ model. In the /export/netnumen/ums-server
operationmodel/log/, for example, opmode directory, execute the shutdown-console.
lserver-20110720-1532-00141.log, search sh command, and then the ums-server\utils\
for the key words read operationtree file, operationmodel\operationmodel.sh command.
and find the NE command packets loaded Wait for some time because all the NEs must be

Figure 1. Current NE Command Packet Loaded by the EMS Version

 Maintenance Experience
www.zte.com.cn

loaded. The old NE modules should be deleted, command in configuration management.


otherwise, the loading will take a long time.
(4) Start the server and reload the SSS model. 3 Lesson Learned
In the /export/netnumen/ums-server directory, 1. Try to recall what you have done
execute the nohup console.sh command to restart before the fault takes place, which
the server. Query the start process by executing the operation may possibly result in this fault,
ps -ef|grep zte command. Select the SSS NE on the and then locate the fault quickly.
U31 maintenance terminal, right-click it, and select 2. Be good at analyzing logs, especially
Load Model, as shown in Figure 3. the logs containing ERR and FAIL fields.
6. After loading the model successfully, restart With abundant information in the logs, you
the U31 maintenance terminal, and execute the same can locate the fault easily. ■

Figure 2. Querying Model

Figure 3. Executing”Load Module”

IMS Core Network Products 


May 2012 Issue 271 Maintenance Instances

Failed to Log in Because the Default


USCC Password is Modified
⊙Li Peng / ZTE Corporation

1 Fault Symptom then enter c, finally press the Enter key.


When debugging the USCC, the 2. In the boot menu, run the c command to
engineer cannot log in by the default user modify the enable password option, and modify
name and password. the password to zxr10.
3. Save the modification and restart the USCC
2 Troubleshooting board.
1. Wait for five seconds, press the 4. You can log in to the USCC by using the
Enter key to stop automatic startup, and password zxr10. ■

IMS Processes 200 Telephone Problems


⊙Li Peng / ZTE Corporation

1 Fault Symptom
This function modifies the DAS of command to query the Centrex configuration
the Centrex to 0. Run the SHOW CTX before the modification, as shown in Table 1.

Table 1. Centrex Configuration Before Modification

CENTREX Calling Centrex Called Centrex Outgoing dial


Name Centrex Type
ID DAS DAS prefix

1379 1501 1100 0 Xx holiday hotel Common Centrex


1510 1510 1100 0 Xx detention house iVPN Centrex
1316 1503 1100 9 Xx coporation Centrex Common Centrex
25 1504 1100 7 Xx coporation Centrex Common Centrex
19 1505 1100 0 xx VPN Common Centrex
20 1521 1100 9 Xx coporation Centrex Common Centrex

10 Maintenance Experience
www.zte.com.cn

Refer to Table 2 for the Centrex configurations after modification.


Table 2. Centrex Configuration After Modification

CENTREX Calling Centrex Called Centrex Outgoing


Name Centrex Type
ID DAS DAS dial prefix
1379 0 0 0 Xx holiday hotel Common Centrex
1510 1510 0 0 Xx detention house iVPN Centrex
1316 0 0 9 Xx coporation Centrex Common Centrex
25 0 0 7 Xx coporation Centrex Common Centrex
19 0 0 0 xx VPN Common Centrex
20 0 0 9 Xx coporation Centrex Common Centrex

After modification, some Centrex short number cannot be connected. These numbers are
started with 2 or 8. From the common office code analysis, the engineer can find that local
numbers of this area are started with 2 and 8. You should perform number transform to such
prefix. During short number processing, number transform must be performed in advance, and
then use the transformed number to query the relation table between Centrex short number and
long number. Number match is failed and an empty number is reported.
Refer to Table 3 for the local SSS office code before modification.

Table 3. Local SSS Office Code Before Modification

Max. Digit Min. Digit Number


Analyzer Analysis High-fee Number Length
Length of Digit Length of Digit Transform
Entrance Prefix Flag Range Match
Stream Stream Index
1014 1 31 3 0 OFF Not matched
1014 2 31 3 103 OFF Not matched
1014 3 31 3 0 OFF Not matched
1014 4 31 3 0 OFF Not matched
1014 5 31 3 0 OFF Not matched
1014 6 31 3 0 OFF Not matched
1014 7 31 3 0 OFF Not matched
1014 8 31 3 103 OFF Not matched
1014 9 31 3 0 OFF Not matched

To solve this problem, you should modify the SSS common office code configuration, which
means to modify digit length from 3-32 to 7-32, and match the digit length. This solution is used
because the incoming calls and local calls of the numbers longer than the short number are
normal. Refer to Table 4 for the local office code configuration after modification.

Table 4. Local Office Code Configuration After Modification


Max. Digit Min. Digit
Analyzer Analysis Number High-fee Number Length
Length of Digit Length of Digit
Entrance Prefix Transform Index Flag Range Match
Stream Stream
1014 1 31 3 0 OFF Not matched
1014 2 31 7 103 OFF Matched
1014 3 31 3 0 OFF Not matched
1014 4 31 3 0 OFF Not matched
1014 5 31 3 0 OFF Not matched
1014 6 31 3 0 OFF Not matched
1014 7 31 3 0 OFF Not matched
1014 8 31 7 103 OFF Matched
1014 9 31 3 0 OFF Not matched

IMS Core Network Products 11


May 2012 Issue 271 Maintenance Instances

After upgrade, there are 200 calls, complaining that all calls are prompted to be empty
numbers.

2 Troubleshooting
After receiving 200 calls, you may remember the digit length modification of the common office
code analysis started with 2 and 8, and the digit lengths of 200 and 201 calls are under 7.
After communicating with the carriers about the fault causes, performing signaling trace and
viewing the configurations, the carriers agree to add the 200 number analysis configuration.
Through configuring 20 prefix analysis from 20-19 and 80-89, the fault is solved, see Table 5.

Table 5. Adding Digit Analysis Configuration

Max. Digit Min. Digit Number


Analyzer Analysis High-fee Number Length
Length of Digit Length of Digit Transform
Entrance Prefix Flag Range Match
Stream Stream Index
1014 20 1 6 3 0 Matched
1014 21 1 6 3 0 Matched
1014 22 1 6 3 0 Matched
1014 23 1 6 3 0 Matched
1014 24 1 6 3 0 Matched
1014 25 1 6 3 0 Matched
1014 26 1 6 3 0 Matched
1014 27 1 6 3 0 Matched
1014 28 1 6 3 0 Matched
1014 29 1 6 3 0 Matched

3 Further Analysis
1. For the short numbers starting with
0 in some Centrex, modify the common
office code 0 prefix (toll), otherwise, empty
numbers will be reported.
2. The 200 prefix fails to match bit
length during SSS digit analysis, but the
200 call is reported to be empty number,
refer to Figure 1.
According to normal flow, although the
bit length fails to be matched, the calls in
sip user=phone mode still will be routed to
MGCFS. The Tel digit analysis of SCSCF
is shown in Table 6. Figure 1. Empty Number is Reported for 200 Call

Table 6. Tel Digit Analysis of SCSCF

Route Selector Unified Resource ID Service Tyoe DNS/ENUM Query HSS Query

2 fj.ctcims.cn Other domain Query DNS Not query

12 Maintenance Experience
www.zte.com.cn

For the URI analysis whose service type is configure Centrex DAS for SSS, and then
other domain, user=phone dos not take affect, and enter Centrex intra-group short number
thus calls are routed to ICSCF according to domain analysis from this DAS. The engineer cannot
name. Only when the service type is local domain, configure a DAS for each Centrex because
SCSCF will route calls to MGCF after tel digit the number of Centrex may be tremendous.
analysis, refer to Figure 2. As a result, it is recommended to share
From above flow, you can find the 200 call a DAS by different Centrex for ease of
failure is caused by three conditions as follows: Centrex migaration and configuration. The
1. The format of AGCF outgoing call is sip engineer does not have to configure DAS for
user=phone. Centrex service implementation, instead, the
2. Due to the modification of SSS digit analysis engineer only needs to scan the relation list
bit length, the 200 number failes to be matched. of long and short numbers of this Centrex
3. In the SCSCF tel digit analysis, the analysis From above analysis, the engineer
property of the prefix 2 is other domain. should handle related NEs comprehensively
As normal configuration, you do not have to for digit analysis and route. Espacially the
modify condition 1. number processing of SSS and CSCF:
Reconfigure condition 2 because it is modified 1. Consider local number, toll number
temporarily and no scenario is considered. and Centrex short number configurations
Condition 3 is not configured properly, and its during SSS digit analysis, especially digit
URI analysis should be modified to local domain. analyaisis of the called numbers.
2. Pay much attention to the differences
4 Lesson Learned among local domain, other domain and local
Before the end of 2010, the engineer must in CSCF digit analysis. ■

Figure 2. Local Domain Flow

IMS Core Network Products 13


May 2012 Issue 271 Maintenance Instances

SCSCF Fails to Route Numbers in


user=phone Format
⊙Ge Huaxiang / ZTE Corporation

1 Fault Symptom
The CSCF routes numbers in the calls by host part without judging whether
user=phone format as numbers in tel “user=phone” is carried. As a result, route results
format. If a call (sip:999777@ims.tisca of the SIP platform do not agree with the expected
li.net;user=phone) is accessed, the results.
S-CSCF transfers messages to tel:999777 Firstly, the service module routes the numbers
for analysis and to VMS, but actually, it in tel format based on 999777, and then obtains
transfers messages to the I-CSCF. correct route set which will be handled by the SIP.
Because the request-uri is not the tel format, the
2 Troubleshooting SIP platform will perform another route analysis
During route analysis, the service without judging “user=phone”. Instead, it uses the
module performs TEL digit analysis domain name ims.tiscali.net for number routing
according to user part, and obtains route directly. For the S-CSCF, the outgoing domain
set. In the request sent to the SIP platform, name analysis is sent to the I-CSCF, and thus
the request URI in SIP URI format still messages are sent to the I-CSCF finally. I-CSCF
carries the “user=phone” . The SIP finds that 999777 does not exist, and returns 404.
platform judges the request URI format. If Solution one: The calling SSS regulates
the format is TEL, the SIP platform routes numbers, sends numbers in “user=phone” format of
the calls by the route set of services. If the caller to the calling S-CSCF after caller triggering,
format is SIP URI format, the SIP platform and the format changes to tel. Refer to Figure 1 for
performs route analysis again, and routes default policies of SSS number regulation.

Figure 1. Default Policies of SSS Number Regulation

14 Maintenance Experience
www.zte.com.cn

When regulation is required, associate CSCF Solution one is recommended.


node ID and this policy, as shown in Figure 2.
Solution two: When analyzing 999777, the 3 Lesson Learned
S-CSCF inserts “top route”, and then the SIP This instance helps you to learn the
platform will analyze the number based on the interior route processing mechanism of the
correct “top route” instead of the domain name of service module and the SIP platform. ■
the “user=phone” number.

Figure 2. Associating CSCF Node ID and Policy

IMS Core Network Products 15


May 2012 Issue 271 Maintenance Instances

Unregistered Callee Gave No Reply


⊙Ge Huaxiang / ZTE Corporation

1 Fault Symptom 2 Analysis


The callee subscribes to the called If the callee does not subscribe to the called

unregistered triggering template. The unregistered triggering SP, the I-CSCF rejects the

S-CSCF sends SAR messages to the HSS call after interacting with HSS LIR/LIA. The HSS

for subscription data downloading, but the returns failure codes in the DIAMETER messages,

HSS makes no reply to SAR messages, see Figure 1.

and the flow is failed after timeout. If a user subscribes to the called unregistered

Figure 1. HSS Returns Failure Codes in the DIAMETER Messages

16 Maintenance Experience
www.zte.com.cn

triggering SP, the HSS returns Diameter information be sent to the HSS to download user data.
in the DIAMETER messages, see Figure 2. The HSS makes no response because the
After receiving LIA messages, the I-CSCF SAR messages contain no host name of
transfers SIP messages to the S-CSCF. Because the opposite end. The HSS fails to process
the user is not registered and the S-CSCF does such SAR messages.
not save related information, SAR messages must

Figure 2. HSS Returns Diameter Information

IMS Core Network Products 17


May 2012 Issue 271 Maintenance Instances

Check SAR requests sent from the handles the calling/called unregistered SP differently.
S-CSCF and make sure there is no HSS If the caller/callee does not registered, the
host name, see Figure 3. S-CSCF sends SAR to the HSS to download user
data during a call. At this time, the UC module
3 Solution invokes digit analysis by using a routing selector,
Add HSS host name in corresponding and analyzes callee/caller numbers. Nevertheless,
routing analysis. The I-CSCF analyzer is the domain selector and the routing analysis of this
used in this route analysis, but the CSCF domain selector together form the routing analysis

Figure 3. Checking SAR Requests

18 Maintenance Experience
www.zte.com.cn

which is used by the calling S-CSCF for transferring the flow recovers, see Figure 4.
messages to the I-CSCF. As a result, the HSS host
name and domain name must be configured on this 4 Lesson Learned
analysis. The HSS host name corresponds to the A common route analysis is performed
Destination host name in DIAMETER messages, based on entrance selector and top route/
and the Home domain name corresponds to request line. The UC module is a little
Destination realm name. After configuration, the different, so the engineer needs to pay
DIAMETER messages carry HSS host name, and much attention to the difference. ■

Figure 4. Detailed Decoding

IMS Core Network Products 19


May 2012 Issue 271 Maintenance Instances

Registration Flow Terminal Fails to


Subscribe to UP10
⊙Ge Huaxiang / ZTE Corporation

1 Fault Symptom 2 Cause Analysis


After a soft-terminal is registered It won’t be the user subscription data or user
successfully, the Wiphone soft terminal data problem because the user is the same. When
originates service subscription failure comparing two registration flows on different
messages to UP10. The same user terminals, the engineer finds the problem.
uses SoftDA to register the subscription There are two ASs, one is SSS, and the other
successfully. is UP10. For third party registration, ASs will be
triggered according to triggering priority. SSS takes
priority of UP10 in subscription, so the S-CSCF
triggers SSS and then triggers UP10. The process
from successful registration to UP10 third-party
registration, S-CSCF sends messages to SSS,
SSS returns 200 OK, and finally S-CSCF sends
messages to UP10. Here it is assumed that the
entire process takes 0.5s, if terminal sends a
message to UP10 in 0.5s, a problem is raised
because UP10 has not implemented third-party
registration.
Before sending registration, the SoftDA
Figure 1. SoftDA Registration Flow
interacts HTTP with UP10, so the time of
sending subscription is always longer than 0.5s.
Nevertheless, the Wiphone has no interaction, and
the subscription is sent within 0.5s, as a result, the
subscription is failed.
See Figure 1 for SoftDA registration flow.
See Figure 2 for Wiphone registration flow.

3 Solution
Because the UP10 performs third-party
registration after the SSS, the SSS third-party

Figure 2. Wiphone Registration Flow


registration interaction flow costs most of time.

20 Maintenance Experience
www.zte.com.cn

After adjusting third-party registration priority, the


problem is solved and UP10 third-party will be
originated first
See Figure 3 for the adjusted registration flow.

4 Lesson Learned
If a user subscribes to several third-party
registrations, the registrations will be performed
according to priority. All ASs operate properly. The
next AS will be triggered after the first AS 200 OK
message is received.
Figure 3. Adjusted Registration Flow
In case of SESSION_ CONTINUED, the flow
will continue no matter what failure response the In case of SESSION_ TERMINATED,
AS returns or the AS makes no response. The the flow is stopped no matter what failure
next AS will be triggered if there is a third-party response AS returns, or the AS makes no
registration, if not, the flow is ended and the user response. The S-CSCF will send user logout
will not be logged out. and trigger third-party logout request. ■

Failed to Call User=phone Number


⊙Zhao Yanhua / ZTE Corporation

1 Fault Symptom
For all the end-to-end calls generated by AGCF, transferred by calling S-CSCF, an error
calling numbers are in siptel format (user=phone). occurs during outgoing route analysis. See
After the I-CSCF receives INVITE request Figure 1 for flow tracing.

Figure 1. Flow Tracing

IMS Core Network Products 21


May 2012 Issue 271 Maintenance Instances

2 Analysis configuration instructions, if you select SIPtel number


Based on user=phone function routing in SIP format, only common routing analysis
configuration, select global soft parameter policy can be performed, such as setting the route set
28 on I-CSCF, set the 22th digit to “1”, and to other domain NEs. Other policies are unavailable,
the 23th digit to 0. See Figure 2 for switch such as HSS query. Nevertheless, the next-hop of
definition and value range. this route analysis should be called S-CSCF. This
Enable failure observation on the route analysis handles all HSS query functions based
NMS. You cannot locate problems based on domain name.
on URI configuration analysis, because
local domain analysis of the domain 3 Solution
name is configured properly. Although To avoid above faults, during user=phone
route analysis is reported in failure, local outgoing route analysis, you can analyze SIPtel
domain name analysis is configured number in TEL format, configure related digit
in URI configuration. As a result, you analysis in TEL digit analysis, configure HSS query
cannot locate fault causes based on route function, and finally get corresponding home hosts
analysis. of the callees. The flow is processed properly after
Enable the full flow trace function on configurations, and the calls can be made and
the NMS, see Figure 3. Through this flow, released properly.
you find that you fail to get route set if the
outgoing route analysis is performed based 4 Lesson Learned
on the domain name of the SIP number Master NE functions and how to implement
converted from SIPTEL number. After functions logically, and then you can locate the root
further analyzing the user=phone functional causes promptly. ■

Figure 2. User=phone Switch Selection and Configuration

Figure 3. Full Flow Trace

22 Maintenance Experience
www.zte.com.cn

TD Mobile Phone Fails to Dial IMS Video


⊙Zhang Heng / ZTE Corporation

1 Fault Symptom
The TD mobile phone fails to dial IMS video of Dezhou, empty number is prompted.

2 Troubleshooting
1. When performing tracing, S-CSCF finds that the area code of Qingdao, rather than Dezhou,
is added before the message from VIG to SCSCF. As a result, CSCF replies a 404 message and
there is a problem occurs in VIG processing, see Figure 1.

Figure 1. INVITE Message

2. Check VIG configurations, and find that only format. In conclusion, perform the following
the area code of Qingdao is configured for digit operations on VIG:
analysis. Based on on-site conditions, for an incoming When receiving 0086, delete 00 and add +
call, MGCF sends BICC messages to VIG in 0086/86 before the number. When receiving 86, add +
format. VIG sends SIP messages to SCSCF in +86 to the number directly as shown below:

number-operation create 1 operation remove-called digits 2 position 1


number-operation create 2 operation insert-called digits + position 1

number-operation-set create 1 number-operation-ids 1 2 name 3G_0086_IMS


number-operation-set create 2 number-operation-ids 2 name 3G_86_IMS

number-manipulation create 1 stage pre match-incoming-protocol bicc match-


called-number 0086[0-9]* number-operation-set-id 1
number-manipulation create 2 stage pre match-incoming-protocol bicc match-
called-number 86[0-9]* number-operation-set-id 2

3. For outgoing calls, VIG receives SIP When receiving +, delete + and add 00
messages from S-CSCF in +86 format. VIG sends before the number, as shown below:
BICC messages to MGCF in 0086 format.

number-operation create 3 operation remove-called digits 1 position 1


number-operation create 4 operation insert-called digits 00 position 1

number-operation-set create 3 number-operation-ids 3 4 name IMS_3G_0086

number-manipulation create 3 stage pre match-incoming-protocol sip match-


called-number [+]86[0-9]* number-operation-set-id 3

IMS Core Network Products 23


May 2012 Issue 271 Maintenance Instances

4. View the VIG digit analysis as shown below:

admin/config> show config number-analysis

match match match


incoming calling called selection route
id protocol number number priority id
-- -------- ------- ------ --------- -----
1 sip 100 1
2 bicc 100 11

admin/config> show config number-operation-set

id name number operation ids


-- ---- --------------------
1 3G_0086_IMS 1 2
2 3G_86_IMS 2
3 IMS_3G_0086 3 4

admin/config> show config number-operation

id operation digits position noa


-- --------- ------ -------- ---
1 remove called 2 1
2 insert called + 1
3 remove called 1 1
4 insert called 00 1

admin/config> show config number-manipulation


number
match match operation
incoming called selection set
id stage protocol number priority id
-- ----- -------- ------ --------- ---------
1 pre bicc 0086[0-9]* 100 1
2 pre bicc 86[0-9]* 100 2
3 pre sip [+]86[0-9]* 100 3

5. Redial the IMS user by a TD mobile S-CSCF carries area code. When querying messages
phone, and the call still fails. Through from gateway office to MGCF, the engineer finds they
signaling trace, the engineer finds the called carry area code, see Figure 2. Therefore, the fault
number from MGCF to VIG carries no area may be caused by called number properties, VIG
code, but the called number from MGCF to digit analysis and trunk configurations.

Figure 2. Messages from Gateway to MGCF

24 Maintenance Experience
www.zte.com.cn

6. This fault has nothing with the called number


properties because both the called number carrying
no area code and that carrying area code are
user numbers. Check the VIG digit analysis on the
MGCF and find no number transform is performed,
as a result, this fault must be related to VIG trunk
configurations, see Figure 3.
7. Check VIG trunk configurations and trunk
group flags on the MGCF. When comparing it with
SCSCF configurations, the engineer finds that the
InterNum flag is missing in the VIG trunk group,
which means that the called number in international
format is used. Add the trunk flag InterNum,
see Figure 4, and then the MGCF can carry the
flag normally. The command is SET TGFLG:
TG=3,ENABLE= “InterNum”.
8. Nevertheless, an empty number is still
prompted when you dial the video. After performing
signaling trace, the ISCSCF returns a 404
message, see Figure 5.
9. At this time, the I-CSCF has queried S-CSCF
server list to the HSS, and the LIA message is
queried successfully and the SCSCF1 domain
name is returned, see Figure 6. Figure 6. SCSCF1 Domain Name is Returned

Figure 3. Comparing Called Number Properties

Figure 4. Adding Trunk Group Flag InterNum

Figure 5. ICSCF Returns 404 Message

IMS Core Network Products 25


May 2012 Issue 271 Maintenance Instances

10. Perform SCSCF1 URI analysis 3 Lesson Learned


based on VIG entrance. If a 404 message is 1. If the call is failed, trace messages to check
returned, the data may be wrong. Query the which NE returns the release message.
SCSCF1 URI analysis of the VIG entrance 2. VIG digit analysis configuration is too simple,
and find it is not configured. Configure the try to implement digit analysis on the MGCF. The
URI analysis and the fault is removed. VIG analyzes and transforms digits between BICC
and SIP.
ADD URIANA:ID=231,NAME=
3. If MGCF carries no area code or digit
“QDA_VIG_JN_SCSCF1_ALL”
analysis is performed, check called/calling number
,RTSEL=190,URI=“scscf1.
sd.chinamobile.com”,METHOD= pre-analysis, digit analysis data, and query the
“ACK”&”BYE”&”CANCEL”&” properties of the called number. Besides, check
INFO”&”INVITE”&”MESSAGE”&” various flags in the trunk group.
NOTIFY”&”OPTIONS”&”PRACK”&” 4. For the errors returned by ISCSCF, locate
PUBLISH”&”REFER”&”REGISTER”&”
the exact call flow of the errors. For digit analysis,
SUBSCRIBE”&”UPDATE”,SRVTYPE=
check TEL digit analysis and URI digit analysis. ■
“HOMEDOMAIN”,SRVATTR=
“COMM”,DNSENUMQRY=“DNS”;

26 Maintenance Experience
www.zte.com.cn

503 Error is Return


⊙Yu Jiang / ZTE Corporation

1 Fault Symptom find the ENUM query is performed on the


When configuring the ENUM query during digit S-CSCF, see Figure 2.
analysis, the terminal makes a call, and the S-CSCF Click GETALIVEDNSLINK interface
returns a 503 error. messages. You fail to obtain the dnslink(
returned retcode is a non-zero value, see
2 Troubleshooting Figure 3.
1. Capture packets on the terminal. The terminal 3. Query the DNS link status by
sends an INVITE message to the core network, and executing the SHOW DNS LINK STATUS
then the core network returns a 503 error. Based on command. The first DNS link is unreachable,
the x-zte-cause head in the 503 message, this error see Figure 4.
is caused by the S-CSCF (host name is scscf3.pplat. 4. Start the link tracing and trace No.1
com.cn), see Figure 1. DNS link, and then execute the DNS link
2. Start the Full-Trace on the S-CSCF. You can continuity command, see Figure 5.

IMS Core Network Products 27


May 2012 Issue 271 Maintenance Instances

Figure 1. x-zte-cause Header

Figure 2. Start Full Trace on the S-CSCF

Figure 3. Failed to Obtain dnslink

Figure 4. Unreachable DNS Link

Figure 5. Start Link Tracing

28 Maintenance Experience
www.zte.com.cn

5. Through link tracing, you can find that the cause carried by the error response
10.5.16.103 address (S-CSCF) sends NAPTR from the core network. Start the failure
query to the 10.5.12.188 address (dns server). observation and full trace on this NE to
After the DNS server replies, the NMS returns a locate the exact flow. In this case, the fault
message dns link is reachable!, see Figure 6. is caused by ENUM query. After querying
If the NMS makes no reply after the continuity related interface messages, you find
command is executed, you should configure that the fault is caused by no available
related records on the DNS server. DNS link. To solve the unreachable DNS
6. Re-query the DNS link status, and find it link fault, start the DNS link tracing and
turns to be reachable, see Figure 7. Re-make a execute DNS link continuity command.
call, and the call is successful. Trough tracing and MML command, you
find the CSCF is sending DNS query to
3 Lesson Learned the DNS server. After receiving reply, the
You can locate the faulty NE based on x-zte- CSCF changes the DNS link status to be
available. The fault is removed. ■

Figure 6. DNS Link Continuity Command

Figure 7. DNS Link Status Turns to Be Reachable

IMS Core Network Products 29


May 2012 Issue 271 Maintenance Instances

LINUX Establishes a Mounting Partition


Larger Than 2T
⊙Zhang Wei / ZTE Corporation

1 Fault Symptom
The GPT partition table, also called the GUID
During the charging system debugging,
partition table, supports up to 18EB (E:exabytes,
the CG disk array (ZTE S260 array) should
1EB=2(10) PB=1024PB, 1PB=2(10) TB=1024TB).
be larger than 2T based on customer
Each disk supports up to 128 partitions.
requests, but the MBR partition table used
As a result, you should change the S260
by Fdisk of the LINUX operating system
disk array from MBR format to GPT format by
supports 2T partition at most.
performing:

2 Troubleshooting 1. Unmounting the disk array from the CG Linux


system.
The partition table used by Fdisk is
2. Executing the following commands on two
MBR (Master Boot Record). The 0 panel, 0
CG blades:
channel, and 1 sector are the master boot
sectors. The Fdisk writes 512 bytes here, parted /dev/sdb
also called the MBR record. mklabel gpt
mklabel msdos
The MBR partition table supports up to
mkpart primary 0% 100%
2TB volume, and up to four main partitions
quit
or three main partitions and one expansion
partition. The expansion partition supports 3. Formatting the disk array by executing the
unlimited logic driver. mkfs.ext3 -F /dev/sdb1 command. ■

30 Maintenance Experience
www.zte.com.cn

SIP Error Code 422


⊙Yu Jiang / ZTE Corporation

1 Fault Symptom
The terminal sends an INVITE call request to seconds, see Figure 2.
the IMS NE, and a 422 error code is returned. 3. Query the code stream which traces
INVITE, and you find that the terminal
2 Troubleshooting supports the session timer and the carried
1. Capture packets on the terminal. The terminal session-expires value is 90, see Figure 3.
sends an INVITE message to the core network, and 4. Stipulated by the RFC4028, if the
the core network returns a 422 error code. proxy receives a session update request
As seen from the x-zte-cause head carried in which contains supported head and
the 422 error code, this fault is caused by the No. session-expires head, and the head values
2 instance of No.1 SIP module of the P-CSCF (The are timer, the session update request can
host name is pcscf3.pplat. com.cn). It generates be rejected by 422 response when time
8F00 AC87:Tu session timer the value in intervals in the session-expires head is
session-expires less than Min-SE of config, see shorter than the minimum session time
Figure 1. intervals defined in local policy. Currently,
2. Query the P-CSCF session timer two request methods are available: INVITE
configuration by executing the SHOW SIP and UPDATE.
GLOSESSIONTIMER command, and you find As seen from Figure 3, the terminal
that the P-CSCF session refresh mode is enabled, carries the supported: timer and session-
and the minimum session refresh duration is 1800 expires headers, and the session-expires

Figure 1. Querying the x-zte-cause Header

Figure 2. Query P-CSCF Session Timer Configuration

IMS Core Network Products 31


May 2012 Issue 271 Maintenance Instances

Figure 3. Querying the Code Stream

value is smaller than the minimum time refresh configuration, and modify the session
interval (1800s) configured locally. As refresh mode to NOSESS by executing SET
a result, the Proxy(P-CSCF) return 422 SIP GLOSESSIONTIMER:REFMODE=
response which complies with protocol. “NOSESS”;.
5. There are five handling methods as
follows: 3 Lesson Learned
(1) Query the P-CSCF session refresh If receiving the 422 error response, the
configuration, and modify the minimum service flow checks the x-zte-cause header
session intervals to be smaller or equal by capturing packets in 422 response, quickly
to the session-expires value carried locate the faulty NE and service module type,.
by the request. In this case, modify Query the session refresh timer configuration.
1800s to 90s by executing SET SIP And then modify the minimum time interval to a
GLOSESSIONTIMER:MINREFTIME=90;. proper value. Alternatively, you can disable the
(2) Modify the session-expires value to session timer function or modify the session-
a larger value. expires carried by the code stream sent from
(3) Query the P-CSCF session the terminal. ■

32 Maintenance Experience
www.zte.com.cn

The PP Board Fails to Load Version


Properly
⊙Zhang Peng / ZTE Corporation

1 Fault Symptom
After a 13U shelf is upgraded and the successfully after version loading.
OMPversion is loaded successfully by DTF tool, After the OMP version is loaded
the PP board fails to obtain version from the OMP. successfully, connect it to the fore-end
The PP board keeps restarting. OMP board by ssh, execute the SCS
ShowMcmInfo() command on fore-end
2 Troubleshooting sh 0 to check the process poweron status,
1. Confirm whether the OMP is powered on see Figure 1.

Figure 1. Querying Process Poweron Status

IMS Core Network Products 33


May 2012 Issue 271 Maintenance Instances

After the command is executed EXCLOG_rack_shelf_slot_CPU.TXT


successfully, you find that all process Check PP logs, you find that the board is an
status is work. As a result, the fault is not island and OMP cannot be found, see Figure 3.
caused by OMP poweron failure. May be Instead of fault reasons, the PP board cannot be
the PP board is faulty. started because it cannot connect to OMP. The
2. Check the PP board. BASE panel network interfaces of OMP and PP
(1) Enter exit at the admin prompt are deployed on BASE panels of different GESs.
of the OMP board, and enter su at the $ Therefore, the OMP board and the PP board cannot
prompt. connect to each other if connection between the
(2) At the # prompt, ping the 128.129 BASE panels on two GES boards is interrupted. As
IP address of the PP board. a result, the PP board fails to download version from
The address cannot be pinged the OMP board through BASE panels.
successfully. May be the PP board is 4. Solution: query BASE connection between
damaged or the base cable between PP GES boards. Check whether the cables or network
board and OMP board is damaged, and thus interfaces are abnormal. Check the adjustment.
the PP board fails to download versions from
the OMP board, see Figure 2. 3 Lesson Learned
3. Query related information of OMP Once faults are detected, remove the faults
and PP in the No. 3 directory of the OMP one by one. In case of power-on problem on
board as follows: the bottom layer, check the logs on boards,
● Abnormal record: /IDE0/ and locate fault causes by log print. Beside, the
EXC_OMP.TXT/IDE0/EXC_PP.TXT network solution should be made in accordance
● Abnormal log: /IDE0/EXCINFO/ with criterion. ■

Figure 2. Pinging the IP Address of the PP Board

Figure 3. Board Island

34 Maintenance Experience
www.zte.com.cn

SSS Modifies Play-Tone by Interior


Failure Code
⊙Ge Huaxiang / ZTE Corporation

1 Fault Symptom (1) Flow trace by the NMS


SSS supports too many services. Every Because there are too many flow
scenario corresponds to an independent failure tracing messages, it is inefficient if you
code. Use an unreachable callee for example, the try to find messages row by row. You can
corresponding failure code is 2382365478. SSS open flow trace and trace by user first, and
can modify corresponding tones of some failure then find the failure time in trace by user,
flows through using internal failure codes. see Figure 1.
This section describes how to implement these Later in flow trace, you can locate the
operations by locating internal failure codes quickly. failure message quickly based on this time.
For example, the failure time is 58:30.064,
2 Troubleshooting the time that the system generates this
1. There are two methods to locate interior failure is several milliseconds before, in
failure codes: Figure 2, you can find it is 2 miliseconds

Figure 1. Finding Failure Time

Figure 2. Locating the Interior Failure Code of the Unreachable Callee Scenario

IMS Core Network Products 35


May 2012 Issue 271 Maintenance Instances

before, that is 58:30.062. From this failure and disadvantages:


message, we can find the interior failure ● Flow trace on the NMS: comprehensive
code of this unreachable callee scenario is but low efficient.
2382365478, see Figure 2. ● X-ZTE-Cause: quick but cannot feedback
(2) Query by X-ZTE-Cause the failure caused by abnormal version.
Trough wireshark or signaling packet 2. Modify or delete tone playing based on
capture on the NMS, we can find the interior failure code.
X-ZTE-Cause field from the failure (1) The corresponding tone of the interior failure
response reported by NEs. Convert this code is displayed, see Figure 3.
field from hex to decimal notation to (2) There is only one Mo Flag, meaning that the
acquire the failure reasons. This section tone playing function on the callee side is removed
uses the unreachable callee failure as an and only the calling SSS can play tones. You can
example; the X-ZTE-Cause value is as modify Tone ID or delete tone playing based on
follows: actual requirements, see Figure 4. If Mo Flag is set
In X-ZTE-Cause: SSS-CSS:8e000326, to Yes, the calling tone playing is modified, while
SSS-CSS refers to the NE that generates No indicates to modify the called tone playing.
the failure response and a correspond
module, 8e000326 is the interior cause 3 Lesson Learned
value. The decimal notation after The X-ZTE-Cause helps to locate faults
conversion is 2382365478. and learn the interior failure code quickly, which
Two methods have their advantages enhances efficiency greatly. ■

Figure 3. Showing Tone of the Interior Failure Code

Figure 4. Modifying the Tone ID or Deleting Tone Playing

36 Maintenance Experience

You might also like