You are on page 1of 28

Preface Maintenance Experience

Monthly for GSM Products


In this issue of ZTE's "Maintenance Experience", we continue No.38 Issue 117, July/2008
to pass on various field reports and resolutions gathered by ZTE
Engineers and Technicians from around the world.
The content presented in this issue is as below:
Eleven Fault Instances
Maintenance Experience
One Special Document
Editorial Committee
Have you examined your service polices and procedures
lately? Are you confident that your people are using all the tools at Director: Qiu Weizhao
their disposal? Are they trained to analyze each issue in a logical Deputy Director: Chen Jianzhou
manner that provides for less downtime and maximum customer Editors:
service? A close look at the cases reveals how to isolate suspected Jiang Guobing, Zhang Shoukui, Wu Feng,
faulty or mis-configured equipment, and how to solve a problem Yuan Yufeng, Tang Hongxuan, Chen Huachun,
step by step, etc. As success in commissioning and service is Li Gangyi, Gu Yu, Song Jianbo, Tian Jinhua,
usually a mix of both discovery and analysis, consider using this Du Jianli, Qu Ruizheng, Zhang Zhongdong,
type of approach as an example of successful troubleshooting Liu Xianmin, Wang Zhaozheng, Liu Wenjun,
investigations. Wang Yapping, Lei Kun, Wang Tiancheng, Cai
While corporate leaders maintain and grow plans for Hongming
expansion, ZTE employees in all regions carry out with individual
Technical Senior Editors:
efforts towards internationalization of the company. Momentum
Zhang Niantao, Wu Yongjun
continues to build, in all levels, from office interns to veteran
engineers, who work together to bring global focus into their daily Executive Editor:
work. Li Fenglian
If you would like to subscribe to this magazine (electronic
version) or review additional articles and relevant technical
materials concerning ZTE products, please visit the technical
support website of ZTE Corporation (http://ensupport.zte.com.cn). Maintenance Experience
If you have any ideas and suggestions or want to offer your Newsroom
contributions, you can contact us at any time via the following
email: doc@zte.com.cn. Address: ZTE Plaza, Keji Road South, Hi-Tech
Thank you for making ZTE a part of your telecom experience! Industrial Park, Nanshan District,
Shenzhen, P.R.China
Postal code: 518057
Contact: Song Chunping
Maintenance Experience Editorial Committee
Email: doc@zte.com.cn
ZTE Corporation
Tel: +86-755-26770600, 26771195
July, 2008
Fax: +86-755-26772236
Document support mail box: doc@zte.com.cn
Technical support website: http://ensupport.zte.
com.cn
Contents

Fault Instance

Failed to Restrict International Outgoing Call......................................................................................2

MSC Performance Data was Lost.......................................................................................................4

Failed to Enter the Software Management Interface of VCS Dual Servers.........................................5

Frequent Call Drop in one Office.........................................................................................................6

OSB Service Failed.............................................................................................................................7

No Tone on either Side......................................................................................................................10

Wrong Calling Number Displayed in Subscribers Bill....................................................................... 11

Call Failed after Long-time Waiting...................................................................................................12

Failed to Query Bills in Batches........................................................................................................13

Communication Error Codes Appeared Among the Modules............................................................14

Indicators on MSC Alarm Plane Were ON.........................................................................................15

Special Document
How to Expand MSC Single Module to Multi-module........................................................................18
July 2008 Issue 117 Fault Instance

Failed to Restrict International


Outgoing Call
Chen Yu/ZTE Corporation

Symptom local subscriber of one office could also dial the


After the subscriber subscribed the international call with the format 0086, and only
BIOC flag, the engineer found that the the international call with the format 1230086 was
restricted.
When performing the interconnection operation
between the V1 MSC on the site A with the V1
HLR and V3 HLR on the site B, and testing the
international outgoing call restriction, the engineer
found that the result was the same.

Analysis and Solution


1. Subscribe the BIOC (Barring International
Outgoing Call) flag on the HLR agent for the tested
number, as shown in Figure 1.
2. C h e c k t h e M S f l a g m e s s a g e f o r t h e
subscriber who subscribed the flag on the VLR.
The message barOutgoCall=2 means that the
Figure 1. Agent-Basic Service subscription was successful, as shown in Figure 2.

 Maintenance Experience
www.zte.com.cn

3. After the subscriber subscribed the BIOC


flag, the international call would be shifted to the
analyzer. At that time, MSC would query whether
the BIOC flag was subscribed from the VLR and
then decide whether the call was allowed.
On the site, the pre-analysis function was used
to convert the number and to differentiate the
charging rate. In the pre-analysis selector, analyze
the number 0086 and 1230086 as 000386
and 000986, and then enter the 101 originating Figure 2. MSflag Message
number analysis selector. 0003 and 0009 would
be sent to the No. 58 route link (the peer end
judges the charging rate).
When comparing the parameters setting of
these two numbers in each analyzer further, the
engineer found that area attribute of these two
numbers in the pre-analysis was different. The
number 1230086 was analyzed in the entry of
the special number analyzer of the pre-analysis
selector and the Area Attribute is Default
Attribute. The number 0086 was analyzed in the
entry of the long-distance analyzer and the Area
Attribute is International Number.
4. In the international analyzer entry of the
pre-analysis selector, modify the area attribute
of the number 0086 as Default Attribute. After
the modification, the engineer found that the
Figure 3. Analyzer EntryArea Attribute
subscriber under MSC could successfully restrict the
international call when performing the dialing test.

GSM Network Products 


July 2008 Issue 117 Fault Instance

MSC Performance Data was Lost


Sun Jingrun/ZTE Corporation

Symptom 129 and save the modification.


The carrier reported that the 2. Restart the process smopsrv_col.exe on
performance data of MSC between 7:40 the 129 server.
am and 9:10 am, Sept 5th, was lost and 3. Run the file for twenty minutes and
the number of the VLR subscribers was copy the following logs under the directory
also abnormal. zxg10: mopsrv_error.log, mopsrv_info.log,
When querying the performance data mopsrv_error.~log, mopsrv_info.~log, oerrlog.zev,
on 5th September, the engineer found oErrMsg.~ze, SerRun.Log and SerRun.~lo.
that there was no basic traffic data and 4. Under the directory zxg10 of 129 server,
the SMS statistics data between 7:40 am modify ShowAllLog=0 in the file mop_pack.ini ,
and 9:10 am. The number of the VLR and save the modification.
subscribers was abnormal, as shown in 5. Restart the process mopsrv_col.exe on the
Figure 1. 129 server.
6. Extract the following data from the 129
database.

use mopdb
select * from mop_msccpudata
use mopdb
Figure 1. Abnormal VLR Numbers
select * from mop_msccpudatax
where sdate<'2007-09-05 09:30' and
Analysis and Solution sdate>'2007-09-05 07:00'
It was doubted that the abnormal use mopdb
number of the VLR subscribers had select * from mop_msccpudataxx
a relationship with the loss of the where sdate<'2007-09-05 09:30' and
performance data. When checking the sdate>'2007-09-05 07:00'
serrun.log file on 129, the engineer found use mopdb
that the performance statistics process select * from mop_msctrafdatax
restarted continuously between 5:44 and where sdate<'2007-09-05 09:30' and
6:40, and the last restart time was about sdate>'2007-09-05 05:00' and ntype=0
9:10.
use mopdb
Extract the log according to the
select * from mop_msctrafdataxx
following methods:
where sdate<'2007-09-05 09:30' and
1. Modify ShowAllLog=1 in the file
sdate>'2007-09-05 05:00' and ntype=0
mop_pack.ini under the directory zxg10 of

 Maintenance Experience
www.zte.com.cn

The analysis result was as follows: Aiming at the BDE abnormality, the
The foreground buffer could only store 2000 engineer analyzed the operating system
packets. The buffer of the foreground was full, log of the 129 server, located the patch
so the data between 7:35 am and 9:10 am were of SQL Server as SP3 and found that the
lost. After 9:10, the server began to receive the SP4 patch was not installed. After installing
data from the foreground, so these data were the patch SP4 patch program, the problem
recorded. was solved.

Failed to Enter the Software Management


Interface of VCS Dual Servers
Han Jinliang/ZTE Corporation

Symptom configured in the VCS login interface.


Under the normal condition, the VCS would In the host file, it was allowed to
start after the command hagui was inputted. At set many hostnames. But at least one
that time, a new local connection was created. After hostname interconnected to the internal
clicking the OK button, a dialog box appeared. network IP must be consistent with the
Input the user name and the password (admin/ hostname inputted in the VCS login
password) to enter the system. interface.
After creating a local connection and clicking Edit the /etc/host of two servers and
the OK button, the engineer found that a dialog box add the IP address and the hostname for
appeared. The prompt message was as follows: the peer server:
==================================== 127.0.0.1 localhost
VCS ERROR V-16-10-4 129.0.12.131 v440-1 loghost
Could not connect to a live system in the cluster 129.0.12.132 v440-2
n2401:14141 closing all windows.
==================================== Experience Summary
When configuring the file /etc/host,
Analysis and Solution more attention should be paid to the
When analyzing the LOG, the VCS engineer hostname. At least one hostname should
found that the hostname interconnected to the be consistent with the hostname inputted
IP in the /etc/host was not consistent with that in the VCS login interface.

GSM Network Products 


July 2008 Issue 117 Fault Instance

Frequent Call Drop in one Office


Hao Li/ZTE Corporation

Symptom still existed. In order to reduce the effect, it was


In one MSC, the call drop phenomenon required to power off one of the MP for a while.
was frequent. The call between the intra- After that, the call drop phenomenon disappeared.
office subscribes and the outgoing calls 4. After analyzing the alarm log further, the
of the local subscribers had the call drop engineer found many Status check notification
fault. notification messages and the address=DEFFE
message.
Analysis In the control shelf, MP communicated with
1. When tracing the interoffice ISUP other boards in this shelf through different bus
message, the engineer found that MSC addresses. The corresponding addresses of
sent the REL message after the call was different boards were as follows:
connected for several seconds or dozens D0xxx: SMEM board
of seconds, and the call was released D1xxx: the first COMM board
at the same time. The reason of the call DAxxx: the tenth COMM board
release was the network is unavailable or DBxxx: the eleventh COMM board
the network is abnormal and the resource DCxxx: the twelfth COMM board
is unavailable. DDxxx: the thirteenth COMM board
2. When checking the level-1 to DExxx: the fourteenth MONI board
level-3 alarms, the engineer found that the The message Status check notification meant
active/standby MP switched frequently. that the bus check was wrong. It was doubted that
When checking the error. log of MP, the there was something wrong with the MONI board.
engineer found that only the continuous
handover information about MP was Solution
recorded. The MP belongs to PIII, so it After plugging/unplugging the MONI board,
was confirmed that the MP switch was not both MPs worked normally. When performing the
caused by the INT13 interruption. call tracing, the subscriber did not find any call drop
3. After plugging/unplugging and phenomenon and the MP did not switch frequently
replacing the SMEM board, the problem again. The problem was solved.

 Maintenance Experience
www.zte.com.cn

OSB Service Failed


Wang Ke/ZTE Corporation

Symptom the following schemes were given:


At present, the subscriber under the local MSC 1. Under the ODB and OSB service
dialed the long-distance mobile subscriber or the configuration of MSC, restrict the prefixes
PSTN subscriber with the following format 0+ (3, 03, 05, 07 and 09.
5, 7 or 9) + area code + number. When dialing the 2. I n t h e p r e - a n a l y s i s , a d d a n
local subscriber, it was only required to dial seven analyzer entry and then analyze the
digits directly, such as 2XXXXXXX, 3XXXXXXX, numbers of which the prefixes are 03, 05,
4XXXXXXX and 8XXXXXXX. 07 and 09. Add 00 before these prefixes.
The carrier should subscribe the long-distance After that, the numbers are changed to
restriction service for the subscriber. On the site, 0003, 0005, 0007 and 0009. The attribute
the OSB method was adopted. On the HLR, it is Unknown.
was required to subscribe the OSB service for the 3. Under the international analyzer
subscriber. Under the Mobile Number Analysis entry of the MO number analysis selector,
Configuration > ODB and OSB Service Call analyze the number of which the prefixes
Restriction Number of MSC, four long-distance are 03, 05, 07 and 09. Delete the first
prefixes 03, 05, 07 and 09 are configured. In fact, two digits 00 and select Unknown for the
the long-distance calls were not restricted, but the Attribute.
local call of which the prefixes were 3, 5, 7 and 9 This scheme aimed at the subscribers
were restricted. who subscribed the OSB service. Perform
the digit diversion (calling) operation
Analysis through the pre-analysis. Before the format
When checking the number analysis of the number sent by CC to VLRMAP,
configuration of the MSC, the engineer found that two digits 00 are added. After VLRMPA
both the pre-analysis selector and the MO analysis deleted the prefix 00, the number will
selector used the DAS101. In the DAS101, the digit be matched with the prefixes 03, 05, 07
diversion (calling) operation was not done for the and 09 in the OSB restriction list. After
prefixes 03, 05, 07 and 09. All these indicated that the above operations, the long-distance
the number format sent by CC to VLRMAP was function was restricted.
just the 03XXXXX format dialed by the subscriber. For the subscribers who did not
Furthermore, VLRMAP had deleted the digit 0, subscribe the OSB service, it was also
so it does not match with the number in the list of required to perform the above pre-
OSB service restriction. According to the normal analysis flow when dialing these long-
requirement of OSB service, the format sent to distance prefixes. However, after VLRMAP
VLRMAP by CC should be CC+ area code + received the number format, such as
number. MSC needed to dial these prefixes, so 0003, sent from CC, it would analyze the

GSM Network Products 


July 2008 Issue 117 Fault Instance

outgoing format of the 0003 in the MT number analysis 3. Subscribe the OSB service for the subscriber in
selector according to the normal flow if the calling the HLR agent, as shown in Figure 3.
subscriber did not subscribe the OSB service. So, it was
required to configure the prefixes 03, 05, 07 and 09 in
the international analyzer entry and specify them to the
corresponding outgoing route links.

Solution
1. Under the Configuration Management >
Supported Service Option Configuration of HLR, check
the checkbox Support OSB Service, as shown in Figure 1.

Figure 3. OSB Subscription Service

4. Under the Mobile Digit Analysis > ODB&OSB


Service Configuration, configure the restricted digits, as
shown in Figure 4.

Figure 1. Service Configuration Supported by HLR

2. Under the Configuration Management > PLMN


Specific ODB Configuration of HLR, configure the self-
defined OSB service, as shown in Figure 2.

Figure 4. Mobile Number Analysis

5. Under Security Management > VLR Security


Variables, set 2 for Selecting of Call Barring Pattern.
6. In DAS101, select 1 from drop-down list in the
Called New Service Analyzer Entry, as shown in Figure
5. In this analyzer, add and configure 03, 05, 07 and 09;
perform the digit diversion operation for these prefixes;
add 00 before the number; and select Unkonwn for
Figure 2. ODB Configuration Diverted Area Attribute, as shown in Figure 6.

 Maintenance Experience
www.zte.com.cn

7. In the international analyzer entry 4 of DAS101,


configure 03, 05, 07 and 09; perform the Digit Diversion
ID (Called) for these prefixes; delete the first two digits 00;
and ensure that the format of the number in the outgoing
IAM message is 03XXXXX, as shown in Figure 7.

Figure 5. Digit Analysis SelectorDAS 101

Figure 7. Digit Diversion (Called)

After that, all the data modification is completed. At this


time, the long-distance restriction service is normal.

Figure 6. Digit Diversion (Calling)

GSM Network Products 


July 2008 Issue 117 Fault Instance

No Tone on either Side


Feng Dongwu/ZTE Corporation

Symptom no tone, so it was doubted that the problem had a


When using the specified trunk to test relationship with the CIC.
the expanded circuit, the engineer found 3. After adjusting the CICs of these two
that the circuit had bi-directional no tone abnormal circuits, the engineer found that neither
phenomenon. directions of the call had tone.
4. When eliminating the faults of the circuit,
Analysis and Solution the circuit on the site A had the cross-wiring
There were three new circuits that were problem. After adjusting the circuits, the engineer
expanded to the MSC-A, but two of them found that the test result was normal.
had no tone on either side. In general,
this phenomenon had a relationship with Experience Summary
the trunk. Maybe, the trunk circuit was This is a normal problem, and the possible
self-looped or the CIC of the data was reasons are as follows:
inconsistent. 1. In general, the no tone problem results
1. When checking the trunk status in from the abnormal trunk, such as the self-loop
the dynamic management, the engineer trunk or the cross-wiring.
found that the trunk was in the idle status. 2. Maybe, the data configuration of the peer
2. W h e n t r a c i n g t h e s i g n a l i n g end is incomplete. It is allowed to check whether
message, the engineer found that the the signaling has the charging message. If the
calling and the called IAM message were peer end receives the IAM message and sends
normal, and the charging message ANM the REL message directly, it means that the data
existed. Both directions of the call have configuration of the peer end may be wrong.

10 Maintenance Experience
www.zte.com.cn

Wrong Calling Number Displayed in


Subscribers Bill
Xu Gaofeng/ZTE Corporation

Symptom 4. After modifying the name of the file c:/


All the subscribers in the whole network of one zxg10/bill/obtezte.dll under the original charging
office were the prepaid subscribers and the bills server, covering this directory with the file
were sorted on the IN side. At present, the version obtezte.dll in the version V3.03.005 (the current
of MSC was V3.03.004. MSC version is V3.03.004), and restarting the
At present, some subscribers wanted to process of the charging server, the engineer
open the postpaid function, so it was required found that the problem was solved.
to install the bill server to sort the postpaid bill.
After installing the charging server, the engineer Experience Summary
found that calling number in the bill coded from The original CDR view is shown in Figure 1.
the background had the digit 0. For example, Instructions:
the normal number was 9235, but the number
displayed in the bill was 920355.

Analysis and Solution


1. First, it was doubted that there was
something wrong with the charging setting table.
After checking the parameter setting of that table,
no abnormality was found.
2. Then, it was doubted that there was
something wrong with the foreground module. The
number 920355XXXX may appear under the
following two conditions:
i. The ServedMSISDN number of MOC was Figure 1. Original CDR View

00920355XXXX.
ii. The ServedMSISDN number of MOC 1. I t i s a l l o w e d t o c l i c k t h e b u t t o n
was 920355XXXX. The number attribute was 1 Directories to set the path of the original
(international). path. The directory of both QueryResults and
3. When performing the query and analyzer Trace cannot be set through the interface.
tool of the original bill, the engineer did not find any QueryResults is used to save the query result.
TRACE result under the corresponding directory. Trace is used to save the result of the trace bill.
The format of the number in the original bill was These two directories only can be set through
correct, which meant that the foreground module modifying the configuration file obcdrman.ini.
did not have the number problem. The incorrect 2. Save the query result to the directory
codec program of the background resulted in the QueryResults. The file name includes the date
incorrect format of the calling number in the bill. and the time.

GSM Network Products 11


July 2008 Issue 117 Fault Instance

Call Failed after Long-time Waiting


Chen Yunlong/ZTE Corporation

Symptom
The client reported that the call failed after a
long time waiting when the MSC was busy.

Figure 1. Signaling Trace (calling)


Analysis and Solution
1. When tracing the signaling of the calling
subscriber at the port A, the engineer found that
MSC sent back the DISCONNECT message to
BSC, as shown in Figure 1. The failure reason was
normal,unspecified, as shown in Figure 2.
2. The paging times set in MSC security
variable was 2. The value in Figure 3 minus 1 was
the times of re-paging.
3. W h e n t r a c i n g t h e s i g n a l i n g f o r t h e
called party, the engineer found that two paging
Figure 2. Failure Reason Value
messages were sent by MSC (the second paging
was sent three seconds after the first paging), as
shown in Figure 4. The MSC did not receive the
Figure 3. Security Variable Value paging response message after ten seconds, so it
sent the DISCONNECT message to BSC.
4. The MSC paging mainly aimed at LAC.
After clicking the mcaPAGING_M message, the
engineer could see that LAC was 3E9, as shown
in Figure 5 and Figure 6. This LAC was attached
with three BSCs, so three PAGING messages
could be seen.
Figure 4. Signaling Trace (called)

Figure 5. mcaPAGING_M Message(1)

12 Maintenance Experience
www.zte.com.cn

Experience Summary
From the signaling tracing, it was known that
the wireless side did not give response to the
PAGING message sent by MSC, which resulted
in the call failure. The failure reason needed to
be checked by the wireless side and it had no
relationship with the MSC side.
In additional, the time interval between the first Figure 6. mcaPAGING_M Message(2)
paging message and the second paging message
was three seconds in the old version of MSC. A MSC V3.04.12.P21.B1, and it could be used to
new timer T169 was added in the new version modify the time interval of the paging.

Failed to Query Bills in Batches


Chen Yu/ZTE Corporation

Symptom
When querying the bill of a whole day from the
charging management system on the charging
client in batches, the carrier found that the capacity
of the bill query system was restricted (the default
value was 5M), as shown in Figure 1. At this
time, the bill of a whole day could only be queried
section by section according to the time. The
carrier required to expand the upper limit capacity
Figure 1. Warning
of the bill.

Analysis and Solution


The field in the file was as follows:
The file c:\zxg10\bill\querycfg.ini on the client
//the max length of the decode files(KB)
had the related configuration options, and they
[Que_MaxSizeDecodeFiles]
could be used to adjust the upper limit value. If the
Value=5000
upper limit value was too large, the query may take
a very long time. On the site, the engineer could On the site, the capacity of the bill was only
modify it to a proper value according to the real 35M, so the changed value 50000(50M) could
conditions. meet the requirement.

GSM Network Products 13


July 2008 Issue 117 Fault Instance

Communication Error Codes Displayed


Among the Modules
Chen Yu/ZTE Corporation

Symptom module 4 and module 1 had the error message


On the site, when the single module Modules communication message has error code
of MSC was expanded to more modules, continuously. The phenomenon resulted from
the engineer found that the peripheral the MPMP (the communication board among the
modules) board of these two modules.

Analysis and Solution


The alarm appeared with a certain frequency.
On each board, this alarm appeared every fifteen
minutes (few alarms appeared every twenty
minutes or thirty minutes). Module 3 had no alarm.
When checking the status of the MPMP board
according to the tool Moniproj.exe, the engineer
found that the alarm display had a relationship
with the received error record, as shown in Figure
1. When a new error record generated, an alarm
would appear after a period of time (such as fifteen
minutes).
Figure 1. MPMP Board Status When trying to adjust the software and

14 Maintenance Experience
www.zte.com.cn

exchange the communication port 1 and the 4 with that of module 3, and check whether
communication port 2 of the adjacent office any error code appears (within fifteen
connected to module 3 and module 4, the problem minutes or more) with the tool Moniproj.
still existed. When replacing the MPMP boards for exe.
module 1 and module 4 with that of module 3, the In the end, the engineer found that
problem still existed. All the above meant that the two DDSN boards on module 4 resulted
problem had no relationship with module 1. When in the alarm. If these two network boards
checking the connection of the cable, the engineer worked as the active board, the alarm
also did not find the problem. would generate on the corresponding
In the physical configuration, when adjusting module. The error code received from the
the HW time delay on module 4, the engineer MPMP board of the module could be seen
found that the problem still existed. Note: it was through the tool Moniproj.exe. When these
only required to adjust the HW occupied by the two network boards worked as the standby
related board, the 4~19 HW and the 68~83 HW board, this alarm would not appear.
could be adjusted. It was not required to adjust the The problem could be solved after the
HW corresponding to the DT; otherwise, the DT fault network board was replaced.
board would not be started.
The DDSN/DSNI/FBI board also had a Note: When replacing the network
relationship with the communication of the board /SDNI/FBI at night, maybe E1
Modules. module 4 was a new expanded module, will lose synchronization, the link will
so it was doubted that there was something wrong be broken, which will further have an
with the board on module 4. influence on the service.
Replace the DDSN/DSNI/FBI boards on module

Indicators on MSC Alarm Plate Were ON


Chen Yu/ZTE Corporation

Symptom Analysis and Solution


When expanding the MSC on the site, the In general, if the background 129
temporary 129 server was replaced. After the server was not connected, all the
expansion, the engineer found that all the services indicators on the alarm plate would be ON.
were normal but all the indicators on the alarm When the alarm plate was connected to
plate were ON. the switch, it would communicate with the

GSM Network Products 15


July 2008 Issue 117 Fault Instance

server. When installing the fault management module,


When installing the background the installation program will display the prompt
program on the client, the prompt message Whether to configure multi-exchange
message whether to control the alarm alarm plane after the prompt message whether
plane appeared. If you select YES, the to control the alarm plane appears. At this time,
server would establish the connection with select YES.
the alarm plate. When the installation program needs to input
If the alarm plate was not installed the exchange number of the installation exchange,
during the installation, but it was required please input Y. This is the number that is allocated
to install the alarm plate, you could modify to the alarm plate.
the information of the registry. If it is required to change the exchange number
The steps are as follows: Y after the background program is installed, you
1. Open the registry editor, and then can modify the file omccfg.ini under the directory
modify the option of the Alarm Plate under ZXG10:
the directory HKEY_LOCAL_MACHINE\
SOFTWARE\ZXJ10B\ALARM\. If the [AlarmPlate]
value is 1, it means the alarm disk is Office=Y
controlled. If the value is 0, it means the
alarm is not controlled. Note: Y is the exchange number of the alarm
2. After the registry is modified, plate.
restart the alarm service process on the
server. Note: After modifying the configuration file,
it is required to restart the alarm service process.
Experience Summary
1. The 129 server connects with After installing the alarm plate, it is required
the alarm plate through the following two to allocate an IP address for the alarm plate. The
methods: details are as follows:
i. Connect the 129 server with the If the IP address of the server is X1. X2. X3.
alarm plate through HUB. Under these X4 (this IP address should be configured for the
conditions, the alarm plate uses the default first network card of the server), it is also required
IP address allocated by the program, and to allocate an IP address X1. X2. Y. X4 for
it is not required to configure the alarm the second network card that is connected with
plate. the alarm plate directly. In which, X1. X2. X4
ii. Connect the 129 server with the correspond to that of the server. Y is the exchange
alarm plate through double network cards. number allocated to the alarm plate.
That is to say, use the crossover network After the alarm plate is configured successfully,
cable to connect the alarm plate with you can open the background process oaSrv.exe
the second network card of the server. and check whether the connection with the alarm is
At this time, it is required to allocate an successful.
exchange number Y for the alarm plate. When double clicking the Alarm Plane Status,
The exchange number Y could be set after you can see the status of the current alarm plane.
the server is installed. If the status is abnormal, you can check whether

16 Maintenance Experience
www.zte.com.cn

the corresponding configuration was normal. three sections of the IP to be configured


2. Configuring the remote alarm plate. should be consistent with that of the
In the latest version of the background, the second network card (the network card
system provides the function of the remote alarm connecting to the alarm plate) in the O&M
plate. That is to say, the alarm plate could be agent, and the fourth section could be
configured on the client or the remote O&M agent configured at random.
in order to facilitate the maintainers. After the remote alarm plate is
During the remote alarm plate configuration, configured, the server also could control its
the connection mode between the alarm plate own alarm plate. These two alarm plates
and the client or the remote O&M agent is just the have no influence on each other when
same as that described in the above context. But it they run.
is required to allocate an IP address for the alarm 3. Normal Problems.
plate on the O&M agent no matter whether the During the usage of the alarm plate,
alarm plate is connected to the O&M agent through the alarm plate often had the reset
the HUB or the dual network cards. phenomenon, which caused that five
In the main interface of the fault management indicators on the alarm plate were ON.
system, after selecting the menu Setting > Other This phenomenon resulted from the
settings, the corresponding dialog appears. abnormal alarm process on the server.
When configuring the remote alarm plate, it is The abnormal process of the alarm
required to check Remote client attaches alarm service was caused by many reasons.
plate first, and then allocate an IP address for the In general, the bad performance of the
alarm plate. database could result in this phenomenon.
If the remote alarm plate is connected to the alarm This phenomenon was improved greatly
plate through HUB, the IP address configuration of after the alarm program was optimized. In
this alarm plate is very flexible. This IP address only additional, the abnormal modules, such as
needs to be in the network section of any IP address the performance statistics and the network
configured in the PC of the alarm client. management of OMC could damage BDE,
If the remote alarm plate is connected to the which may further result in the abnormality
alarm plate through dual network cards, the first of the alarm server.

GSM Network Products 17


July 2008 Issue 117 Special Document

How to Expand MSC Single Module to


Multi-module
Chen Yu / ZTE Corporation

1 General Planning check, data backup, installation and debugging of


The reconstruction from a single the hardware and the data configuration.
module to multi-module includes three ii. Cutover before the reconstruction and
phases: during the normal reconstruction.
i. The preparations before the iii. Adjustment after the cutover.
reconstruction include the hardware The architectures before and after the
reconstruction are shown in Figure 1.

2 Preparations before
Reconstruction
2.1 Checking running status before the
reconstruction
(1) MSC/VLR system check
1) Check the historical alarm records to see
whether there are some serious faults in the recent
two weeks.
Figure 1. Architecture Instance 2) Check the performance statistics to see

18 Maintenance Experience
www.zte.com.cn

whether the traffic statistics in the recent week is ii. For the original module 2, put
normal. the reconstructed FBI board and the
3) Check whether the parameter setting of the COMM (MP level) board to the slot of the
SQL Server on the O&M server and the charging corresponding shelf, but do not insert them
server meets the requirement. (that is to say, do not power on them).
4) Check the running status of the charging Connect the fiber to the correct location
server. Check whether the charging bills of the and label it clearly.
current day received from the foreground and iii. During the reconstruction, the new
sent to the charging center are normal. Check the HUB should be close to the old HUB
file oblog.log and the file obaecur.log under the in order to facilitate the network cable
directory C:\ZXG10\TRACE on the charging server modification. After the reconstruction,
131 and 132 to see whether they have some the adjustment operation is easy and the
abnormalities. processing flow meets the requirement.
5) Check the running log of each module MP 3) Prepare the clock plan
from the foreground to see whether there are some For the clock adjustment, it is required
abnormal running records in the recent two weeks. to prepare the plan before the expansion
If the device has no essential problems, it is and the reconstruction, such as, the
allowed to arrange the reconstruction. Otherwise, it board adjustment, the cable layout. If this
is required to handle the hidden troubles before the operation is neglected, the expansion and
reconstruction. the reconstruction time will be delayed, the
(2) Check the hardware of the reconstructed clock adjustment will fail or the expansion
device and the cutover will fail.
1) Check hardware device i. For the BITS clock used by the
First, confirm whether the hardware required original single module, it is required to
by the reconstruction is completely prepared. The adjust the CKI board from the original
hardware includes: all boards of the new rack, the single module to the center module during
added FBI board of the old rack, COMM board, the expansion and the cutover, and it
the optic fiber used to connect the center rack and is also required to lay the cables to the
the original single module, the HUB for debugging, center module in advance. At this time, it
one PC for debugging, the version software of the is required to extract the clock from BITS.
current network and the software system to be It should be noted that the imported BITS
installed. clock should have 2-channel input at
On the site, if the MP model is inconsistent (for least. It is allowed to select 2Mb and 2Mhz
example, the original 2# module is PII MP, but the according to the details of the carrier. It
new module is PIII MP), it is required to unify the is required to pay more attention to the
MP model to ensure that the MP model of each impedance match of CKI board. The
module is consistent. unmatched impedance will degrade the
2) Check hardware connection configuration BITS clock of the system and will further
i. For the fiber connection between the affect the steady running of the system.
periphery module and the center module, the fiber ii. For the E8k clock used by the
from each periphery module to the center module original single module, the clock
should be on a pair of CFBI board separately. of the original module 2 is the E8k

GSM Network Products 19


July 2008 Issue 117 Special Document

clock extracted from E1 by the SYCK (save it as the text file). You can check the file
board. During the expansion and the through the tool OBDBTOOL.
reconstruction, it is required to perform In the basic configuration management system,
the following adjustment in that evening. back up the configuration data (note: adopts the
The adjustment steps are as follows: the BCP method and the SQL normal method).
SYCK board of the center module (the Record the value of the variables in the variable
new module 2) extracts the clock from E1, management system (you can save all the values
and then module 2 distributes the clock or the changed values, and check whether the
through the CFBI board, the FBI board of value in the generated file is consistent with that in
each MPM module and the SYCK board of the variable management system).
each MPM. Record the parameters of the timer. You can
4) MSC system backup save all the values or the changed values.
Back up the original ZXG10 directory Record the data configuration of the link and
of the O&M server to the disk D. the trunk group from MSC system to each office
Back up the directory C:\ZXG10 of direction.
the active/standby charging server to It is required to install the 129 server again
the directory D:\ of the active /standby because of the capacity expansion. At this time,
charging server. the engineer should check whether there is
Back up the important database on the the tone script file msctone.sql used during the
129 server commissioning on the site.
Basic configuration database If the performance statistic has the congestion
ZDB_ZXJ10 setting, it is required to record the statistics
Performance statistics database items that need to be enabled or disabled by the
MOPDB performance statistics. The engineer should pay
Alarm database MOAMSSDB more attention to the collection items that need to
Back up the charging database be enabled by the carrier, such as, each circuit in
MOBMSCVLRDB and record the the circuit group statistics and the office direction in
parameter values in the table SETTING the office direction data, as shown in Figure 2.
Record the setting of the stream control
(execute c:\zxg10\loadctl.exe on OMM server).

2.2 Installing and debugging the


reconstruction device
(1) Debug Process
1) Debug MP.
2) Check the board.
3) Check and update the MP board and the
PP board.
(2) Install the new expanded system
1) Install the new 129 O&M server.
Use one PC as the temporary 129 server of
Figure 2. Circuit Group Statistics MSC. The data, such as, the office number, should

20 Maintenance Experience
www.zte.com.cn

be the same as that of the current 129 server. making the MP:
Install the version of the current office. If there is no After the MP is made, put a clear label
new PC and no other PC to replace the temporary on MP in order to facilitate the later
129 server, the MSC O&M agent could be used. debugging.
2) Network connection. The running mode of MP meets the
Module 1, module 2, module 4 of the requirement.
foreground and the new background 129 server 4) Configuring data on the new O&M
comprise a new MSC through a new HUB. Both server.
the network cables and the HUB should be close to The engineer could configure all data
the rack of the server in order to facilitate the wiring again or restore the data to the new O&M
after the normal reconstruction. server and then modify the corresponding
3) Make the MP for module 2, module 2, data. Perform the following steps to modify
module 3 and module 4. the restored data.
During the real reconstruction, module 1, Reserve the exchange configuration,
module 2 and module 4 of the foreground are new, the number analysis, the mobile
and module 3 is reconstructed from the original number analysis, SCCP configuration
module 2. The configuration file omc.cfg under the and LAI configuration.
directory CONFIG used by module 1, module 2 and Delete all trunk configurations and the
module 4 should be consistent with that in module 2. MTP configuration. It is required to
But, the option Version= in the file Version.cfg is record all original configurations before
configured according to the module. the deletion.
Module 1: Version=MSC&&VLR Delete all physical configurations
Module 2: Version=MSC and then perform the corresponding
Module 4: Version=MSC configuration according to the board
Whether the file tcpip.cfg adopts the Ethernet to slot of the real rack, such as slot 1, slot
communicate depends on the setting of the site. 2, slot 3 and slot 4.
The version of the MP under the directory Because a center module and a
VERSION of each new module is the same as that peripheral module are added, so the
of the original module 2. In general, the PP version attribute of the original module 2 is
of module 4 is the same as that of the original changed. At this time, it is required to
module 2, but it is also required to check whether delete the original physical configuration
the PP boards of these two modules are the same. and perform the physical configuration
It is required to check the PP version of module 2 again. It is required to delete all trunk
according to the real board. circuits and the signaling links in all
Use one MP of the module 4 as one MP of configurations before deleting the physical
module 3. During the reconstruction, the single configuration; otherwise, it is not allowed
MP of both module 3 and module 4 run normally. to delete the physical configuration.
After the successful reconstruction, use one MP of Configure the physical data again
module 2 as the standby MP of module 3, and use and connect the networking. During the
the other MP of module 2 as the standby MP of communication configuration among the
module 4. modules, the number of the communication
The following must be paid attention to during module used by the peripheral module is

GSM Network Products 21


July 2008 Issue 117 Special Document

2. The HW time slot on the board of the new After the debugging, connect the fiber of both
module 3 should be consistent with that of the original module 2 and the new module 2. At this
the current network. time, it is not required to power on the FBI board
Configure the capacity of the office again. on the original module 2. Connect the fiber of both
Configure the capacity for VLR again module 4 and the new module 2.
(apply for the password to modify the 6) Tone adjustment.
capacity of VLR), and put VLR on The temporary 129 server is new, so it is
module 1. required to re-load and adjust the tone script
Configuring the signaling link and the msctone.sql. After performing this script in the
trunk data again. script analyzer, it is required to check whether the
This configuration should be consistent service tone edit item and each KB edit item are
with that of module 2. That is to say, change consistent with that in the tone table maintenance
the number of the module from 2 to 3, but service of the original 129 server; otherwise, the
the other data should be consistent. tone play after the cutover will fail.
In the basic data configuration, the In the tone table management, it is also
number of the module that belongs to required to re-select or add the service tone type
the BSC configuration is 2. After the for the tone board unit; otherwise, the tone cannot
adjustment, it is required to modify the be played after the cutover.
number as 3. 7) Modify the security variables and the timer
5) Debugging. according to the backup security variables and
It is required to debug the data after the the timer. After the modification, it is required to
data configuration. Use the three racks of synchronize them with the foreground (When using
module 1, module 2 and module 3 to perform one MP of module 4 as one MP of the module 3, it
the debugging. is also required to synchronize the security variable
Put the MP of module 3 and module 4 and the timer on the rack 4).
on module 4. At the same time, adjust the 8) Adjust the performance statistics items
location and the fiber connection of the and the alarm filtering items on the temporary 129
corresponding CFBI board. Adjust other server and keep them consistent with the original
boards according to the conditions on the 129 server.
site. Try to simulate the environment of 9) Adjust the stream control parameters and
module 3 and module 4 according to the keep them consistent with that on the original 129
rack of module 4. Synchronize the data of all server (with the tool loadctl.exe).
tables and perform the test.
Use the tool to check whether the 3 Reconstruction process
resource in each module or among the 3.1 Reconstruction steps
modules is normal. Use the tool MONIPROJ 1) Generate the bill from the foreground by
and the diagnostic and test program to manual, and start the 130 charging program from
check whether the communication among dual servers (viz, power off the charging process of
module 1, module 2, module 3 and module 4 the background).
is normal. In the fault management system, 2) Power off the active/standby MP of the
observe whether the board of module 1, original module 2, and unplug the active/standby
module 2, module 3 and module 4 is normal. MP. Replace them with the debugged module 3.

22 Maintenance Experience
www.zte.com.cn

Do not power on them. new 129 server to the directory C:\ZXG10 of


3) Unplug the network cables of the original the active/standby charging server. Start the
129 server (all the cables of the dual network cards charging server progress of from the dual
should be unplugged), unplug the network cables servers and check the generation of the bill.
of the temporary 129 server from the new HUB and 12) Switch the dual charging servers
plug them into the HUB of the original 129 server. and check whether they run normally and
4) Unplug the MP dual network cables from the connection of them is normal after the
the new HUB of module 1, module 2 and module configuration is modified.
3 and plug them into the original HUB according to 13) Use one MP of the original module
the former planning. 2 as the standby MP of module 3 (reserve
5) Move the CKI board of the original module another MP of the original module 2). When
2 to the new module 2. Module 3 and module 4 testing module 3, the engineer found that the
extract the clock from FBI board. active/standby MP runs normally after the
6) Power on the FBI board and the COMM switching.
board (MPMP) of the original module 2 (the new 14) Check the charging bill with the
module 3). charging center.
7) Module 3 starts the MP and synchronizes
all tables again. It restarts the MP and re- 3.2 Changeback steps
synchronizes all the tables. Each restart operation If it is required to change back the
is done after the MP starts the alarm. When it is operation for some unsolved reasons,
required to restart the VLR module and there is a perform the following steps:
PPS subscriber, it is allowed to restart the active 1) Start the charging server progress of
MP after the synchronization alarm of the active/ the 130 server from the dual servers.
standby module disappears (VLR subscriber data 2) Power off the active/standby MP of
synchronization). module 3 and unplug the MP. Replace it with
8) Synchronize the security variables of all the active MP of the original module 2. Do
the modules and verify them. Check whether the not power on it.
timer is set just the same as before. 3) Unplug the network cable of the
9) Use the test tool to check whether the temporary 129 server and plug the network
resource of module 3 is normal. Use the tool cable of the original 129 server into the
MONOPROJ and the routine test program among original interface. Unplug the dual-network
the modules to check whether the communication cables of module 1, module 2 and module
among module 3 is normal. In the fault 4 from the HUB. Unplug the CKI board
management system, observe whether the board inserted in module 2 and plug it into module
of module 1 to module 4 is normal. 3. Adjust the clock to recovery the clock
10) Perform each kind of service test and extraction of module 3.
record the result (check whether the voice plays 4) Unplug the FBI board and the
normally). MPMP board from module 3.
11) After the call test is confirmed, back up the 5) Restart the MP of the original
file omccfg.ini under the directory C:\ZXG10 of the module 2.
active/standby charging server, and then copy the 6) Resynchronize all the tables and
file omccfg.ini under the directory C:\ZXG10 of the restart the MP; and then resynchronize all

GSM Network Products 23


July 2008 Issue 117 Special Document

the tables again. for the original 129 server again. It is also required
7) R e s y n c h r o n i z e t h e s e c u r i t y to create the performance statistic database again
variables. because of the change of the module number.
8) Perform the dialing test and record Otherwise, the database will run abnormally. It is
the result. only required to reserve the original databases
9) Restore the file omccfg.ini under for other systems instead of creating them again.
the directory C:\ZXG10 of the original After the installation, restart the system and
active/standby charging server. Start the check whether each system runs normally (check
process of the 130 charging server from whether the tone Flash file used on the site exists
the resource of the dual servers and check under the directory C:\ZXG10). Adjust the value of
the generation of the bill. the performance statistics database as the original
value, create the CSM module, and open the
4 Adjustment after cutover collection items and the congestion parameters of
After the formal reconstruction, it is the performance statistics.
required to observe the system for one Back up the data on the new 129 server as
week in order to check whether the data the SQL file and restore the data on the original
configuration is correct and whether the 129 server. The original 129 stream control
device runs normally. If the device runs parameters (loadctl.exe) should only includes the
steadily, cooperate with the carrier to configuration of module 2. It is required to delete
perform the following adjustment in order the configuration of this module and then re-
to complete this expansion reconstruction. configure the stream control parameters of the
Adjust the ON/OFF switch on the periphery module 3 and module 4.
reserved MP of the original module 2 and After verifying the configuration data, unplug
use the MP as the MP of module 4. the network cable from the HUB of the new 129
Delete the data under the directory server. Connect the network cable of the original
\DATA of MP, delete the file TIMER.CFG 129 server to the current HUB.
under the directory \CONFIG, and modify It is required to adjust the traffic load of the
the setting of the file \CONFIG\VERSION. outgoing trunk and the signaling link among
CFG. Synchronize all the tables and re- each module according to the planning and the
synchronize them after the system restart. requirement of the carrier. After the single module
Test the active/standby switching after the is expanded to multi-module of the center rack,
active/standby MP runs normally. all the traffic is also on the module 3 (the original
Set the other MP of the original module module 2). At this time, it is required to allocate
2 as the standby MP of module 3. the traffic of module 3 to module 3 and module 4
Back up some important databases regularly.
of the original 129 server, such as the
performance statistics database, the alarm
Note: The trunk wire of the backplane on
database, and the authority database. Back
module 3 should be long enough in order to be
up all files under the directory C:\ZXG10.
connected to module 3 or module 4. Try to distribute
When the original 129 network card
the signaling link and the traffic of each office to
performs the self-loop operation, it is
module 3 and module 4 regularly.
required to install the background software

24 Maintenance Experience
www.zte.com.cn

GSM Network Products 25

You might also like