Vous êtes sur la page 1sur 48

Preface

In this issue of ZTE's "Maintenance Experience", we continue to pass


on various field reports and resolutions gathered by ZTE Engineers and Maintenance Experience
Technicians from around the world. Editorial Committee
The content presented in this issue is as below:
32 fault instances
Director: Zhou Susu
8 frequently asked questions (FAQs)
Have you examined your service polices and procedures lately? Are Deputy Director: Chen Jianzhou
you confident that your people are using all the tools at their disposal? Are
they trained to analyze each issue in a logical manner that provides for Editor-in-Chief: Yang Cheng

less downtime and maximum customer service? A close look at the cases
Editors:
reveals how to isolate suspected faulty or mis-configured equipment, and
Jiang Guobing, Wang Yaping, Ba Zexue,
how to solve a problem step by step, etc. As success in commissioning
Zhang Shoukui, Wu Feng, Yuan Yufeng, Tang
and service is usually a mix of both discovery and analysis, consider
Hongxuan, Chen Huachun, Ding Guixiang,
using this type of approach as an example of successful troubleshooting Gu Yu, Tian Jinhua, Zhu Wensheng, Ling
investigations. Changwen, Zhang Zhongdong, Liu Xianmin,
While corporate leaders maintain and grow plans for expansion, Wa n g Z h a o z h e n g , C h e n Ta i m i n g , Z h a n g
ZTE employees in all regions carry out with individual efforts towards Mingjing, Wang Haidong, Chen Le, Lei Kun,
internationalization of the company. Momentum continues to build, in Wang Tiancheng, Zheng Hongliang, Wang Tao
all levels, from office interns to veteran engineers, who work together to
Technical Senior Editors:
bring global focus into their daily work.
Zhang Niantao, Wu Yongjun
If you would like to subscribe to this magazine (electronic version) or
review additional articles and relevant technical materials concerning ZTE Executive Editors:
products, please visit the technical support website of ZTE Corporation Li Fenglian, Zhang Jinbao
(http://support.zte.com.cn).
If you have any ideas and suggestions or want to offer your
contributions, you can contact us at any time via the following email:
doc@zte.com.cn. Maintenance Experience
Thank you for making ZTE a part of your telecom experience! Newsroom

Maintenance Experience Editorial Committee


ZTE Corporation Address: ZTE Plaza, Keji Road South, Hi-Tech
August, 2007 Industrial Park, Nanshan District,

Shenzhen, P.R.China

Postal code: 518057

Contact: Song Chunping

Email: doc@zte.com.cn

Tel: +86-755-26770600,26771195

Fax: +86-755-26772236
Contents

Fault Instances

Service Interruption between One MSC and One BSC 3


Area Restriction Service Failure under One MSC 4
Signaling Link between MSC and HLR Intermittently Disconnected 5
Switching System Access Success Ratio in MSC Exceeded 100% 8
In the Performance Statistics, the Number of Call Attempts in Office Traffic Was Inconsistent
with that in Circuit Group Traffic 9
Failed to Set PDP Contexts on the HLR Agent in Batches 11
Lots of 166 Failure Causes Occurred in SMSC 12
Low System Access Success Ratio 15
Failure of Independent IP Tone Playing 17
Failed to Send Paged SM 18
Ineffective Special Number 19
Handling the Database Space Alarm from Performance Statistics 20
Handling the Fault of Having GateOut CDR without MO CDR 21
Handling the Fault of R2 Incoming Call after Call Number Upgrade 22
The number of current subscribers in HLR was inconsistent with the number of actual subscribers 23
Some of Intelligent Subscribers Failed to Roam within a Province, or in Another Province or country 24
Configuration of IMSI Number Analysis 26
Some Subscribers Couldnt be Registered in Network after the Cell Restriction Service Is added in HLR 27
Failure to Login Remote HLR Agent 28
HLR Agent Startup Error When Configuring Data 29
Failure in Backing up HLRDB onto Node 129 31
HLR MAP Process Repeatedly Restarting 32
Failure in HLR Setting PDP Contexts in Batches 33
CUG users Unable to Update Location under ZTE MSC 33
Subscriber Couldnt Access the Network After HLR Cutover 35
August 2007 Issue 52 Fault Instances

Contents

Illegible Codes Appears When the Network Management Client Runs HLR Configuration Management 36
Unsuccessful Cross Number Assignment on the two HLR Databases 36
Subscribers Could not Modify Access Mode in HLR 37
VPN Subscribers Could not be the Calling Party When Roaming 38
Echo Appearing for Outgoing Calls 39
Unsuccessful Location Update for Roaming 39
Difficulty in Making Calls at Busy Hours Under an HLR 40
HLR System Error Occurred in Service Subscription on Agent 40

FAQ

How to Analyze the Location Update Faults 42


How to Handle Unclear Blocked Signaling Links 42
How to Set the Dual Seizure Handling Mode 43
How to Transfer the Database 43
How to Recover the Database in the Suspect Status 44

 Maintenance Experience
Fault Instances www.zte.com.cn

Service Interruption between


One MSC and One BSC
Wang Ke, ZTE Corporation

Symptom
The HW BSC under one MSC2 suddenly
suffered from service failure. The MSC could
receive messages, but failed to send messages.

Analysis
After checking the configuration data at the
MSC side, the engineer found that the arrangement
modes of the SLS in the route chain configuration
from the MSC to other offices were all automatic
arrangement mode, while that configured in the
HW BSC office was manual arrangement mode.
When the fault occurred, the statuses of office
directions and links in the foreground tables
R_Office and R_N7link were normal. One of route
arrangement modes in R_DRout was 0. In normal
situation, no arrangement mode in R_DRout should it was known that the security variable
be 0. in the MSC was 1, and there were 0
If manual arrangement mode is adopted for links. Therefore, the MSC failed to send
the SLS arrangement mode, deleting one link messages.
will cause the 0 link with errors existing in the
dynamic route table. When the office-related or Solution
link-related data are synchronized, or all the tables The problem was solved after the
are transferred, MTP3 will refresh the dynamic engineer changed the SLS arrangement
route table again, and the DB of the module 2 will mode in the HW BSC office to automatic
be synchronized to other modules. arrangement mode, and synchronized the
In the MSC security variable configuration, the data.
default value of Local module preferred during NOTE: During the configuration in
IW/GMSC selecting signaling links is 0. If it is MSC office commissioning, the security
set to 1, the MSC will query the route table when variable Local module preferred during
sending messages. If it is detected that the 0 link IW/GMSC selecting signaling links is set
exists, the MSC cannot send messages. From the to 0 by default. In general cases, do not
feedback from the onsite maintenance personnel, set it to 1.

GSM Network Products 


August 2007 Issue 52 Fault Instances

Area Restriction Service


Failure under One MSC
Wang Ke, ZTE Corporation

Symptom
The area restriction service was enabled under
one MSC, but subscribers in cells subscribing the
service failed to access the network.

Analysis
1. The engineer checked whether security
variables of the MSC were set correctly, and found
that the switches of the following security variables
had been opened.
Cell roaming restriction control
Whether to perform cell restriction during the
Figure 1. Data Configuration In The Location Area location update
Whether to perform cell restriction during the
Mobile Terminated (MT) call
Whether to perform cell restriction during the
short message service
Whether to perform cell restriction during
handoff
Whether to perform cell restriction during the
supplementary service
2. The engineer checked whether the
ZoneCode (as shown in Figure 1) at the MSC side
was consistent with that (as shown in Figure 2) at
the HLR side, and found no problem.
Figure 2. Agent-Area Restriction 3. After checking the capacity of the table

 Maintenance Experience
www.zte.com.cn

R_ZNCD, the engineer found it was big enough


(which was 10000).
4. After checking the foreground with the MSC
probe, the engineer found there were ZoneCodes.
5. The engineer found that Zonecode was not
checked on the VLR Configuration Services
Supported by VLR page, as shown in Figure 3.

Solution
After selecting all the options in the Other
area in Figure 3, the area restriction service was
implemented.

Figure 3. VLR Configuration

Signaling Link between


MSC and HLR Intermittently
Disconnected
Huang Wuxiang, ZTE Corporation

Symptom HLR to the fiber transmission. Then the


The two signaling links between the Miyanddab fiber transmission was exceptional. The
MSC and the Orumieh HLR were intermittently engineer exchanged DTI boards (where
disconnected, so subscribers could update their the two links suffering from intermittent
locations or server as the called party. disconnection were located) of the
MSC, but the fiber transmission was still
Analysis exceptional. So the engineer exchanged
Through inquiry, the engineer knew that this back the DTI boards, and changed
fault occurred at about 9 a.m. To test whether the the transmission back to microwave
new fiber transmission could be normally used, transmission. The two signaling links were
the operator rerouted the microwave transmission always intermittently disconnected.
between the Miyanddab MSC and the Orumieh The engineer connected to the

GSM Network Products 


August 2007 Issue 52 Fault Instances

Miyanddab MSC through Pcanywhere, the BSC and checking the connection of the trunk
and observed these two links dynamically line again, the signaling link still alternated between
in the No.7 link management of the MSC the service status and the non-service status.
dynamic management. The engineer found It was initially doubted that there was something
that one signaling link in Slot 15-Subunit1 wrong with the trunk connection. The engineer
always alternated between the service used the self-loop lines to self-loop these two
status and the non-service status, while signaling links, and found that the same problem
the signaling link in Slot 16-Subunit1 was occurred. Then the engineer tried many methods,
always in the non-service status. At such as reconfiguring the two signaling links after
the same time, the engineer dynamically deleting them, replacing the subunit, replacing the
observed the signaling link in Slot DTI board, replacing the physical slot, switching
15-Subunit2 to the Orumieh EIR, and and restarting the MP, restarting the two signaling
found it was always normal. The signaling boards, and switching the SYCK board. However,
links to other offices or distributed to other the problem still existed.
DTI boards were all normal too. Among the MTP test messages in the No.7 link
The engineer checked the MTP tracing of the signaling tracing, several SLTM and
management configuration and the trunk SLTA messages occasionally occurred, while MTP
configuration. The engineer found that discard error messages always occurred.
CIC 1 of Slot 16-Subunit 1 was allocated The engineer reconfigured a new HLR office
to the signaling link to the HLR, while all and a signaling link, performed self-loop on this
the other time slots and subunits of this signaling link, and rerouted the E1 of one signaling
DTI board had been allocated to the BSC. link of the old HLR office to this signaling link.
After deleting the time slots allocated to This signaling link was in the service status.

 Maintenance Experience
www.zte.com.cn

Among the MTP test messages in the No.7 link variables to reduce the load on the
tracing, SLTM and SLTA messages were normal. signaling link. He/she dynamically
However, once the signaling point code and the GT observed the status of the signaling link,
analysis result of the old HLR office were rerouted and found no intermittent disconnection
to the new HLR office, the new signaling link was problem. The VLR signaling tracing
intermittently disconnected. showed that the location update of the
F i n a l l y, t h e e n g i n e e r d e l e t e d t h e G T subscriber was normal. Late at night when
configuration, the signaling link configuration and the traffic was the lightest, it was observed
the adjacent office configuration of both the old through the VLR probe that the number
HLR office and the new one, retransferred all the of subscribers soared from 3000 to over
tables, reconfigured the old HLR office, and the 10,000 within one minute.
signaling link. The status of the signaling link was After observing from the VLR probe
normal. However, once the GT analysis result was that the speed of subscriber location
added, the problem occurred again. update decreased gradually, the engineer
After the GT analysis result was added, lots of added the subscriber number section
subscriber performed location update and served in E.214 format for the GT analysis
as the called party. Therefore, it was doubted that entrance 1 to allow the subscriber to
lots of subscriber failed to perform location update serve as the called party, and recovered
or serve as the called party for a long time, for the authentication setting in the security
the signaling link had been disconnected for a too variables.
long time. Through the VLR probe, it was known Because the number of subscribers
that about the number of online subscribers had in this office increased quickly, the load
decreased to about 3000 subscribers at about 2 on the signaling link was very heavy,
a.m. Once the signaling link was recovered, lots and the quality of the transmission link
of subscribers would perform location update and was unsteady, it was recommended that
serve as the called party at the same time. The the operator should add the number of
uprush of traffic was easy to cause the signaling signaling links to the HLR.

link congested and blocked. In the No.7 link


observation in the failure observation, congestion
failure messages were found.

Solution
The engineer adopted the method of adding
GT step by step. The engineer first added the
subscriber number section in E.214 format for
the GT analysis entrance 2 for location update of
new subscribers, but did not add the subscriber
number section in E.214 format for the GT analysis
entrance 1 for the time being. So the subscribers
were not allowed to serve as the called party.
The engineer temporarily cancelled
authentication on subscribers in the security

GSM Network Products 


August 2007 Issue 52 Fault Instances

Switching System Access


Success Ratio in MSC
Exceeded 100%
Qin Ruirui, ZTE Corporation

Symptom database MOPDB, setting P7 to -10, BEGIN time


When collecting the network to 0, and END time to 24.
management data in one MSC office, the The engineer executed in the query analyzer:
engineer found that the switching system USE MOPDB
access success ratio exceeded 100%. insert into PM_Trigger (MssId,begindatetime,en
ddatetime,P7)
Analysis values (' ',0,24,-10);
It was initially thought that the switching 2.In the enterprise manager, the engineer
system access success ratio exceeding opened the table PM-TRIGGER in the middle
100% was a false phenomenon. The database MOPDB of the server 129, opened
reason was that the calculation formula the table by right-clicking it, and returned all the
was incorrect, or parameters in some lines to check whether the record in step 1 was
tables in the middle database affected the added.
performance statistics data. 3.After modifying the data, the engineer
checked the network management data within an
Solution hour after 2 hours, and found that the switching
1.The engineer added one record in system access success ratio had decreased to
the table PM-TRIGGER in the middle 100% below.

 Maintenance Experience
www.zte.com.cn

In the Performance Statistics,


the Number of Call Attempts in
Office Traffic Was Inconsistent
with that in Circuit Group Traffic
Guo Yu, ZTE Corporation

Symptom between the office traffic and the circuit


During the network maintenance for one office, group traffic in the signaling flow, the
when performing performance statistics on some statistics data based on these two factors
BSC offices on the MSC, the engineer found were inconsistent.
the number of call attempts counted based on
the office traffic was dramatically different from Solution
that counted based on the circuit group traffic Combining the analysis on the onsite
sometimes. condition, the engineer thought the main
reason why the difference occurred was
Analysis the trunk busy. There was no BSS trunk
The reason was that the recording point of busy item in the office traffic statistics
counting the number of call attempts based on the in the onsite version. The BSS trunk
office traffic was different from that of counting the busy item was added in the subsequent
number of call attempts based on the circuit group V3.04.12.P11.B1 version. Therefore, the
traffic. In normal cases, difference is small. When number of call attempts could only be
serious congestion occurs in the A interface in busy counted based on the basic traffic.
hours, difference is big. For the specific difference, The formula was A=B-C.
refer to the signaling flow in Figure 1. Where, the meanings of A, B and C
In the office traffic, when the MSC/VLR receives were as followings:
the SETUP message from the A interface, one call A: Times of BSS trunk busy in the
attempts will be added. While in the circuit group basic traffic statistics, which was
traffic, when the MSC/VLR receives the assignment counted based on the six call types
completed message from the A interface, one call (international toll automatic outgoing
attempts will be added. call, international toll manual outgoing
Because there were differences in vacant call, national toll automatic outgoing
number, call origination restricted, BSS trunk call, national toll manual outgoing call,
busy, wireless side congested and other aspects local outgoing call, local call).

GSM Network Products 


August 2007 Issue 52 Fault Instances

B: Sum of the number of incoming C: Sum of the number of incoming trunk call
trunk call attempts to each BSC in the attempts to each BSC in the circuit group traffic
office traffic statistics. statistics.

Figure 1. Signaling Flow of the Call

10 Maintenance Experience
www.zte.com.cn

Failed to Set PDP Contexts on


the HLR Agent in Batches
Chen Yunlong, ZTE Corporation

Symptom addresses, while 1 means static


When setting PDP contexts on the HLR addresses)
agent in batches by inputting the text file, the IP address (this field is valid when the
system prompted error occurred when reading PDP address is static, and it is ignored
subscription file, maybe the format is incorrect. when the PDP address is dynamic)
One or more spaces or TABs are
Analysis and Solution needed to separate three fields above.
The onsite HLR version was V3.04.15.P1.B1. According to the format described
When modifying PDP contexts in batches, the above, the engineer inputted subscribers
format of the input text file was composed of: whose PDP context information was to be
Subscriber identity (IMSI or MSISDN) set in the text file, and performed the batch
PDP address type (0 means dynamic operation again.

GSM Network Products 11


August 2007 Issue 52 Fault Instances

Lots of 166 Failure Causes


Occurred in SMSC
Sun Liukui, ZTE Corporation

Symptom CK (*)
The CMPAK SMSC maintenance 11. Servicing MSC or SGSN MS
personnel found lots of 166 failure causes Short Message (GSM 04.11)
occurred in the SMSC, which caused MT 12. MSServicing MSC or SGSN
short message failure. Short Message Acknowledgement (GSM 04.11)
13. Gateway MSC Servicing MSC or SGSN
Analysis MAP_MT_FORWARD_SHORT_MESSAGE_ACK
The signaling flow of the MT short 14. SCGateway MSC
message is as follows: Short Message Acknowledgment (GSM 03.40)
1. SCGateway MSC The MT short message needs to pass through
Short Message (GSM 03.40) the MSC/HLR. According to the above signaling
2. Gateway MSCHLR flow, it is known that the following two cases will
MAP_SEND_ROUTING_INFO_FOR_SM cause MT short message failure.
3. HLRGateway MSC Step 2 in the signaling flow
MAP_SEND_ROUTING_INFO_FOR_ The SMSC asks the HLR for the route
SM_ACK information of the called subscriber. After
4. Gateway MSCServicing MSC or receiving the short message route request from
SGSN the SMSC, the HLR first checks the dynamic
MAP_MT_FORWARD_SHORT_MESSAGE information of the subscriber, checking whether
5. Servicing MSC or SGSNVLR the subscriber is accessible (such as powering
MAP_SEND_INFO_FOR_MT_SMS (*) off the MS), and whether the previous short
6. VLRServicing MSC or SGSN message receiving status of the subscriber
MAP_PAGE/MAP_SEARCH_FOR_M is accessible. If any of these two options has
OBILE_SUBSCRIBER (*) inaccessible mark checked, the HLR will
7. Servicing MSC or SGSNMS directly return absent subscriber to the SMSC,
Page (GSM 04.08) reporting that the MT short message failed. This
8. MSServicing MSC or SGSN point is different from calls.
Page response (GSM 04.08) From step 4 to step 14 in the signaling flow, the
9. Servicing MSC or SGSNVLR MSC may send the MapError message to the
MAP_PROCESS_ACCESS_REQUEST SMSC, prompting that the MT short message
_ACK and MAP_SEARCH_FOR_MOBILE failed.
_SUBSCRIBER_ACK (*) The MT subscribers whom the SMSC perform
10. VLRServicing MSC or SGSN statistics on include those from the local
MAP_SEND_INFO_FOR_MT_SMS_A network and non-local networks. For the

12 Maintenance Experience
www.zte.com.cn

subscribers from non-local networks, the MSC the same time, carrying the cause that is
just performs GT forwarding of the SCCP layer, system error.
but does not perform statistics on whether it is After comparing the call loss with
successful or not. The SMSC is recommended the signaling on the site, the engineer
to perform statistics based on number sections, confirmed that system errors occurred
so that MT short message failure in the local during the MT short message were caused
network and non-local networks can be by NDUB. The MSC call loss and the
counted. VLR call loss were shown in Figure 1 and
After discussing and analyzing the problem with Figure 6 respectively.
the maintenance personnel at the SMSC side, the During the course of MT short
engineer confirmed that the reason why the 166 message, in the case of no response
failure cause occurred was that the MSC sent the to the MSC paging, the MT subscriber
MapError message to the SMSC, whose content powering off the MS, or NDUB, the MT
was system failure. The failure cause code of the short message cannot be completed. But
corresponding VLR call loss was 161. when the MSC successfully pages the
The call loss is described as follows: subscriber, or after the subscriber power
The failure cause code 161 means that the on the MS, the MT short message will be
subscriber being busy is decided by the network completed. In this way, short message
(NDUB). receiving will be delayed.
Involved services: call, short message, and The concept of unsteady status
USSD service. covers the following cases of this problem
Status: when the VLR receives the paging in the protocol:
response that carries subscriber error mNDUB_M No other services can be initiated
during paging the subscriber in the MT call, short during the location update:
message, and USSD service. Between the network and the mobile
Cause: the subscriber is in the unsteady station. The end of the running MM
process during the paging course, such as location specific procedure or the release of all
update, and access. MM connections have to be awaited
When the subscriber is in the steady process before a new MM specific procedure
(such as LU, and USSD service), the VLR MAP can be started (GSM04.08 4.4).
will report the call loss of NDUB, and the MSC If an MM specific procedure is running
will return the MapError message to the SMSC at at the time the request from the

Figure1. MSC Failure Observation

Figure 2. VLR Failure Observation

GSM Network Products 13


August 2007 Issue 52 Fault Instances

CM sublayer is received, and the optimizing the quality of the whole network.
LOCATION UPDATING REQUEST In the light of conditions of the onsite network,
message has been sent, the request the engineer provided other methods to adjust the
will either be rejected or delayed, success ratio of receiving short messages:
depending on implementation, until It was recommended to clear the subscribers
the MM specific procedure is finished who were in inactive status for a long time, so
and provided that the network has not that the ratio of absent subscribers in the HLR
sent a "follow-on proceed" indication, and the related system load could be reduced.
the RR connection is released. (GSM Send group messages only to the activated
04.08 4.5.1.1). subscribers.
No other MM connections are allowed Reduce the resource waste to increase the MT
to be established when the subscriber success ratio.
is in the access phase: For the case of no response to the paging,
When in MM CONNECTION ACTIVE adjust the indoor coverage to increase short
(GROUP TRANSMIT MODE) state, message paging success ratio.
the MM sublayer in the network shall
reject the request for the establishment Summary
of another MM connection by any CM The three main factors affecting the success
layer (GSM 04.08 4.5.1.3.1). ratio of receiving short messages are as follows:
The subscriber being busy is decided by the
Solution network (NDUB).
The system errors occurring when The subscriber powers off the MS (Subscriber
the MT subscriber is in unsteady status Absent for MS turnoff).
are all caused by normal behaviors of the There is no response to paging the subscriber
subscriber. The high NDUB MSC is greatly (Subscriber Absent for Page No response).
related to the quality of the whole network. Therefore, to increase the success ratio of
For example, poor coverage causes lots of receiving short messages, it is necessary to start
location updates. Therefore, the problem from these three factors, and analyze the related
can be thoroughly solved only through call loss according to the above methods.

14 Maintenance Experience
www.zte.com.cn

Low System Access


Success Ratio
Sun Liukui, ZTE Corporation

Symptom When the called MS rings, the calling


The system access success ratio in one office party releases the call before the called
was very low. party answered the call.
3. Equipment Keys Congestion
Analysis The REL message is received from the
Taking the MSC performance statistics in busy opposite end during the local outgoing call.
hours in one month as the reference to perform The failure cause is that the opposite-end
analysis, the engineer concludes that the failure switch is congested.
causes greatly affecting the system access 4. Subscriber Release Early
success ratio are as follows: Before the called party is successfully
1. Mobile Subscriber Unreachable connected, the calling party actively
This cause is that the subscriber cannot be releases the call.
successfully paged or the subscriber powers off the 5. Failures due to Other Reasons
MS. Other causes include the following
2. inging Release Early cases:

GSM Network Products 15


August 2007 Issue 52 Fault Instances

(1) After the called party ringing majority of other causes, while the last two cases
message or backward ACM signaling is are just supplementary to the completeness of the
received, the call fails; and the failure statistics processing.
cause is not Subscriber busy (17) or After analyzing the onsite traffic data, the
Normal release (16). The call is released engineer found that the second case in the other
due to any other causes than Subscriber causes mainly affected the system access success
busy (17) and Normal release (16), ratio in busy hours.
such as no response from the subscriber, 1. The engineer collected the call loss data in
ringing release early, the called party one period of time, which proved this point.
refusing to receive the call (depending on The failure cause occupying the maximum
the release cause inputted by the called proportion was: 0122=Terrestrial circuit unavailable
MS), and outgoing call. because Application for the terrestrial circuit from
(2) The incoming call fails to occupy the A interface or assignment fails. Usually this
the called circuit. occurs in busy hours.
(3) The call origination is restricted 2. This point was proved from the office traffic
in the local office; the release signaling statistics at the same time, as shown in Table 1.
with failure cause value as 31 is received The congestion in the trunk of the onsite network
before the called party in the local office was very serious.
rings or the backward office returns the The traffic per line to N-ABT-B88, N-MDN-B95
ACM message when the called party is and N-PSH-B92 in busy hours was over 1.0elr,
outside the local office during the local while that to N-PSH-B91 and N-MDN-B96 was over
office originating a call (in general, this 0.9elr, which all exceeded the recommend value:
case is less). 0.7elr. Therefore, the serious call loss in busy
(4) Statistics of no statistical item can hours was caused by the congestion in the trunk.
be collected. This case can be ignored.
In general, the first two cases are the Recommended Solution
In the current network, the following two
problems mainly affect the system access ratio:
Table1. Office Traffic Statistics
1. Unsteady transmission causes circuits and
Incoming Outgoing links intermittently disconnected.
Line
Category Seizure Seizure
Time Traffic 2. Serious congestion at the A interface and in
description Traffic Traffic
(ERL) (ERL) (ERL) inter-office circuits causes the network busy.

2007-5-1 17:00 N-ISB-B81 620.69 596.58 0.66


Solving the problems above can greatly

2007-5-1 17:00 N-ISB-B82 422.85 358.89 0.42 improve the system access ratio.
2007-5-1 17:00 N-ISB-B83 486.05 465.45 0.52
2007-5-1 17:00 N-PSH-B91 584.59 552.71 0.99 Summary
2007-5-1 17:00 N-PSH-B92 361.81 341.42 1.03 When processing the system access success
2007-5-1 17:00 N-JHL-B87 260.3 245.95 0.49 ratio problem, first extract some representative
2007-5-1 17:00 N-ABT-B88 496.6 532.56 1 performance data and call loss data, collect
2007-5-1 17:00 N-MDN-B95 344.85 345.82 1.01 statistics on and analyze each kind of failure cause
2007-5-1 17:00 N-ISB-B84 578.37 517.64 0.68 value, find out exceptional cause values, and give
2007-5-1 17:00 N-MDN-B96 308.31 348.67 0.97 the related solution.

16 Maintenance Experience
www.zte.com.cn

Failure of Independent IP
Tone Playing
Chen Yunlong, ZTE Corporation

Symptom was set as 1, which meant no assembly.


The carrier complained that the subscriber with ii. The item of Whether to adjust the
insufficient or 0 balance could not hear the tone independent IP address was set as 1,
playback during a call. which meant that the address format in the
AssistinSSIPRoutingSddr of the ETC would
Analysis be adjusted, with the country code removed
In the onsite network, the tone playback was during the analysis. Setting this item as
independent IP tone playback. The normal flow 0 meant that the number sent by the
is that the SCP sends the ETC operation to the ETC would not be adjusted, which would
MSC. After receiving the ETC operation, the MSC be directly analyzed in the corresponding
originates a call to the IP. However, after tracing call originating selector. Since the item
the signaling, it was found that the MSC did not SPF Combines 3 parameters had been
originate a call to the IP. set as 1, this item should be set as "0",
1. The engineer opened the ETC message sent indicating that the number sent by the ETC
by the SCP to check Assisting SSPIPoutingAddr, would be directly analyzed.
finding that an independent IP address was
entered. The engineer checked the analysis of the Solution
independent IP number, finding that this address The fault reason was that the security
was directly analyzed as the outgoing service variable of Whether to adjust the
[National toll inter-region automatic services]. independent IP address was set as 1.
2. The engineer checked the configuration of After this item was changed to 0, the
the security variables. independent IP tone playback became
i. The option of SPF Combines 3 parameters normal.

GSM Network Products 17


August 2007 Issue 52 Fault Instances

Failed to Send Paged SM


Dai Yi, ZTE Corporation

Symptom after being encapsulated on the bottom layer, the


When the Short Message (SM) was length of two messages Open + FWD will exceed
sent to a certain SMC, it was successful the maximum length of the UDT message. As a
to send one-page SM, but unsuccessful result, the SCCP will divide these two messages
to send two-page SM. The signaling was into the XUDT. If the opposite-end NE does not
shown in Table 1. support the XUDT packet, the service will fail.
By sending the Open request and the FWD SM
Analysis message separately, the occurrence of the XUDT
The MSC processes the SM MO will be reduced, which will improve the compatibility
based on the security variable of SM and ensure the implementation of the service.
Maximum packet size. If the SM length Sending a SM in the paging mode failed after
is less than the security variable value, the Open request was received. The engineer
the MSC will send the three messages found that there was a Provider Malfunction
(Map_Open+FWD SM+delimiter) serially. alarm in the Failure Observation window, that is,
If the SM length exceeds the security the bottom layer considered the SM abnormal.
variable value, the MSC will send the Therefore, it was confirmed that the SMC bottom
Map_Open+Delimeter messages first layer did not support sending SM in the paging
without sending the contents of the SM. mode. The UDT that only contained the transaction
After the SC returns the Open_Cnf and the dialog content (Map_Open+delimiter)
message, the MSC sends the FWD without the component content would be
SM+delimiter messages, sending the SM considered abnormal.
to the SMC. For a SM with over 120 bytes,
Solution
If the opposite-end SMC supports the XUDT,
Table1. Office Traffic Statistics increase the security variable SM Maximum
1 MSCMAP send 8:42 MAP OPEN REQUEST packet size, so that the MSC can send the three
2 MSCMAP send 8:42 MAP DELIMITER REQUEST messages of Map_Open+FWD SM+delimiter
3 MSCMAP receive 8:43 MAP P_ABORT INDICATION serially.

18 Maintenance Experience
www.zte.com.cn

Ineffective Special Number


Huang Wuxiang, ZTE Corporation

Symptom dialed. The engineer first analyzed all


The carrier complained that the special service special service numbers as MSC Special
numbers configured were ineffective. Some Service", and then looked for the reason.
subscribers could not directly dial the special 3. The engineer pre-analyzed the
service number, such as 110, and some calls could special service numbers and the attendant
not be transferred to different attendant consoles console numbers respectively, both getting
based on different cell numbers. the correct analysis results. Therefore, the
number analysis was normal.
Analysis 4. By tracing the messages from the
1. The engineer checked the configuration of A interface, it was found that the network
special number management in the configuration returned the disconnect message
management. Different special service numbers containing the failure cause of incorrect
and different cells had been configured with number format (incomplete number) to
different attendant console numbers. There were the MS. And the same failure message
more than 4000 records in total. was found in the Failure Observation.
2. By querying the number analysis The problem still existed, after changing
corresponding to these numbers, the engineer the values of the longest and the shortest
found that some special service numbers had numbers in the number analysis.
been analyzed as National Interregional Number, 5. The engineer copied the R_MSPEC
and transferred to a specified attendant console table from the MSC foreground to its
number for outgoing after converted at the called background. The engineer opened the
side. And some special service numbers had been MSC probe, and double-clicked the x in
correctly analyzed as MSC Special Service with the top right corner. He/she opened the
the Default Attribute as the number attribute. downloaded R_MSPEC table in the pop-
Therefore, those outgoing numbers after being up window, and found that there were only
converted could be successfully dialed, but could 512 records in this table. However, over
not change different attendant console numbers 4000 records had been configured in the
based on the cell number. However, those numbers special service management configuration
analyzed correctly could not be successfully of this office in fact.

GSM Network Products 19


August 2007 Issue 52 Fault Instances

6. The engineer tried to execute table, the engineer found that the table contained
the dial-up test in the cells where the the failure message of Tables which sending
R_MSPEC table had records, finding that record count bigger than designed capacity.
it was normal. The engineer confirmed Through check, it was found that the
that this fault was due to the inconsistency value of Special Group Table Capacity in the
between the foreground R_MSPEC table office capacity configuration had been set as
and the background configuration. He/she 512. It caused lots of records in the Special
repeatedly started the foreground MP and Service Group Management of the background
then transferred all the tables for three configuration could not be completely transferred
times. But the problem still existed. to the R_MSPEC table of the foreground MSC
module. The engineer changed the value of the
Solution into 4096. After three times of starting the
After checking the Table Capacity foreground MP and transferring the tables, the fault
Check before send log of the transferred was eliminated.

Handling the Database Space


Alarm from Performance
Statistics
Guoyu, ZTE Corporation

Symptom was 1001 M.


The performance statistics gave a After the database size was set consistent
database space alarm, prompting that with its auto-increasing upper limit, the alarm
the database size was less than the disappeared. After some time, the alarm appeared
configured upper limit. again for the database size was restored to 1000
M. After checking the property settings again,
Solution the engineer found that the option of Automatic
After checking the property settings Shrink was selected. After the engineer unchecked
of the SQL performance database, the this option and modified the database size to be
engineer found that the database size was consistent with the increasing upper limit, the
1000 M, and its auto-increasing upper limit problem was solved.

20 Maintenance Experience
www.zte.com.cn

Handling the Fault of Having


GateOut CDR without MO CDR
Hao Li, ZTE Corporation

Symptom Recommended Solution


During the inter-network settlement of the billing It is recommended that the carrier
center, the engineer found that a few records with should discard the CDR whose duration
duration ranging from 1 s to 20 s had the GateOut is less than 1 s without processing. As for
CDRs without the MO CDRs. the CDR whose duration ranges from 1 s
to 20 s, it is also recommended that the
Analysis carrier discard it due to its few quantity
This case for 1-s calls results from the early and inevitability.

release of the calling party. There are only a few


CDRs with the duration ranging from 20 s to 1 s.
The MSC generates the trunk CDRs after receiving
the ANM signaling from the opposite-end office.
The signaling flow is shown as Figure 1. At this
moment, the MSC will report the DP7 to the SCP,
and set 20-s timer to wait for the SCP sending
the continue message to confirm that the call is
put through. The MSC generates the MO CDR till
it receives the continue message. When such
things happens, including the continue message
loss, timer timeout or the subscriber releasing the
call before timeout, the trunk CDR less than 20 s
appears, without the MO CDR. Figure 1. SIGNALING FLOW

GSM Network Products 21


August 2007 Issue 52 Fault Instances

Handling the Fault of R2


Incoming Call after Call
Number Upgrade
Sun Jingrui, ZTE Corporation

Symptom upgrade. Both of them could be used by the


The mobile number of the local subscribers.
network in an office was upgraded from After the number was upgraded, the engineer
095 + SN (six digits) to 095 + SN (seven changed the number stream length of the prefix
digits). Where, 95 was the NDC upgrade 0955X from 9-9 to 9-10, which was in the
rule, that is, 5 was added in front of the analyzer entry 7 of the trunk selector 201 in the
original SN. The new number and the old R2 office. Two days later, the engineer found
number both existed after the number that the new number could not be dialed in the
incoming call from the PSTN (R2). Therefore, the
carrier changed the length of the number stream to
10-10.

Analysis
During a dial-up test, the subscriber
021226526 of the PSTN (R2 office) dialed the
10-digit number (after the upgrade) 0955906834,
but the call was connected to the subscriber
0955590683 (the test number). Shown by the
traced CAS message, the MSC only requested the
9-digit called number 095590683. The system
added 5 behind the called number 095, so the
called number became 0955590683. Therefore,
the subscriber 0955590683 became the called
party. The CAS message traced was shown in
Figure 1 and Figure 2. The digits in the red square
represented the called number 0955590683 that
was received by the MSC.
Different from that of the ISUP, the associated
incoming call number is received one digit by one
digit. For an incoming call home to the local office,
Figure 1. Traced CAS Message 1 whether to stop receiving the number is controlled

22 Maintenance Experience
www.zte.com.cn

by the minimum number stream length that is set in


the number analysis.

Solution
Since the problem occurred only in the R2
office, the engineer sub-divided the trunk selector.
The engineer created the trunk selector 205 in the
ISUP office, changed the length of the number
stream "0955X" to 910 digits, and kept length of
number stream 0955X in the trunk selector 201 of
the R2 office as 10-10 digits.
In this way, the incoming call from the R2
office could dial the new 10-digit number 09555X
upgraded from 0955X. And the old 9-digit number
0955X could not be dialed now. Because there
were less R2 subscribers in the local network, this
method could reduce the influence range to the
uttermost. Figure 2. Traced CAS Message 2

The number of current


subscribers in HLR was
inconsistent with the number
of actual subscribers
Yu Hongliang, ZTE Corporation

Symptom was confirmed after database was queried


The number of subscribers was found abnormal through SQL query analyzer.
when queried through Network Management
(NM) system. An office reported that the number Analysis
of current subscribers in ZTEs HLR, which was 1. The upper NM system was able to
collected by the upper NM system, differed greatly collect data, which accounted for normal
from the number of actual subscribers. The fault connection between area A and area B.

GSM Network Products 23


August 2007 Issue 52 Fault Instances

2. There is an NM system interface collected by the upper NM system.


machine in HLR system, which connects 4. Analysis shows that this data statistic error
to HLR through a switch inward and was caused by fault of operation and maintenance
connects to the upper NM system outward. system.
This interface machine collects HLR data
once every a period of time and stores the Solution
data in its database. The data collected 1. The following script on database was loaded:
by the upper NM system is exactly the declare @retcode int
data stored in the NM system interface begin
machine. The data of current subscribers exec hp_sys_measure @retcode output
are collected by OMM. exec hp_sys_measurecc @retcode output
3. After entering the performance end
statistics system of the client, and then 2. The script was set to run at wee hours every
querying subscribers in HLR Agent day.
System, it was found the data of current Problem was resolved after successful load of
subscribers were exactly the incorrect data scripts.

Some of Intelligent
Subscribers Failed to Roam
within a Province, or in Another
Province or country
Duanmu Xiaoting, ZTE Corporation

Symptom data for intelligent service were deleted on agent.


Some intelligent subscribers failed
to roam within a province, or in another Analysis
province or country. However, they could 1. The ZTE engineers performed analysis on
access the network after their subscription the failure of a VPMN subscriber, who roamed to
some places and could not access the network.
They viewed the location update process of this
subscriber through signaling tracking. The signaling
Figure 1. Signaling Flow flow is shown in Figure 1.

24 Maintenance Experience
www.zte.com.cn

2. It was found that when inserting subscriber constructors (e.g. an empty SEQUENCE
data in the process of location update, the peer- type). That is, the encoding of the content
end VLR returned a MAP_NOTICE to HLR, which of any data value shall consist of zero, one
caused the subscriber information data unable to ore more octets.
be inserted into VLR and thus resulted in failure of 7. However, some VLRs are not
location update. compatible with this kind of condition
3. When the ZTE engineers deleted CAMEL mentioned above. When receiving the
information of this subscriber on the agent, and request for inserting the second package
then viewed the location update flow, it was found of subscriber data, they would consider
that the flow was normal and the subscriber was the code stream of these subscriber data
able to access network. illegal and return a rejection message to
4. The reason why this fault occurs is that the HLR if they find VLR CAMEL flag only but
peer-end VLR returns a rejection message to HLR no such detailed CAMEL data as OCSI
after it has received the request from HLR for in the code stream in the package, thus
inserting subscriber data. resulting in location update failure.
5. TM reports the abnormal structure and
components call loss resulting from cause 112 Solution
after it has received the message, and sends 1. Delete VPLMN information of
MAP_NOTICE_IND (problem diagnosis value is subscribers so that location update may
0) to MAP. Because the subscriber has subscribed succeed.
too much information, the HLR needs to divide 2. Subscribe VPLMN information
the subscriber data into packages when inserting again. When the agent subscribes VPLMN
them. In the process of encoding, at the time of information for the subscriber, HLR only
TM encoding VPMN server information, it is found needs to insert VPLMN subscription
that the length of the current code stream encoded information to VLR, so the inserted
has reached boundary value of the package, so subscriber data is not so large that HLR
some parameters like OCSI are not encoded in needs not to divide them into packages
this package, however, VLR CAMEL flag has when insert them. Therefore, VLR CAMEL
been encoded in this package. This causes, in the flag and OCSI information are in a same
subscriber data inserted in VLR, the first package package when they are inserted in VLR.
has CAMEL data flag but no CAMEL information, The VLR thus consider the code stream
while the CAMEL information is in the second correct and insert these data successfully.
package.
6. It is totally normal when this kind of thing Conclusion
happens during encoding, because VLR CAMEL The reason why the locations of
information is defined as Sequence type in VPLMN subscribers fail to update when
the protocol. The Sequence type may have no they roam to some VLRs is that those
detailed content during encoding, which is clearly VLRs have imperfect compatibility in
explained in MAP protocol, as shown in the code stream. Causes are the same for
following information taken from Section 17.11 of other kinds of subscribers failures of in-
GSM0909-710 protocol: province roaming, inter-province roaming,
There is no restriction to the use of empty or international roaming.

GSM Network Products 25


August 2007 Issue 52 Fault Instances

Configuration of IMSI Number


Analysis
Wang Xiaosong, ZTE Corporation

Symptom is HLRGT; E214 coding is adopted for Result


The GT number of HLR of the Indication is ISDN.
subscribers locations could not be
acquired after the IMSI prefix was directly Solution
analyzed as HLRGT. 1. Acquire ISDN number of HLR.
On the interface of adding or modifying mobile
Analysis digit analysis index, the analyzed result indication
1. IMSI structure analysis is configured as Result indication is HLRGT, which
IMSI structure: MCC+MNC+H0H1H2H3+ transforms the IMSI digit prefix to HLRGT.
XXXXXX, wherein H0H1H2H3 uniquely 2. In Translator dialog box, select ISDN/
identifies one HLR. telephone numbering plan for configuration of
2. Analysis on E164 and E214 coding Number Plan.
methods 3. GT translation data
The E164 code is directly converted Select the above-mentioned GT translator to
into MSISDN number of HLR, or the work out GT translation data toward ISDN number
converted MSISDN number is properly of HLR.
lessened according to the quantity of Because the ISDN number of HLR is unknown
the numbers digits. and only CC+NDC is available, the configuration is
E214 coding method: MCC+MNC (in as following:
IMSI) CC + NDC, with the rest of i. On the interface of adding or modifying mobile
number copied. digit analysis index, the analyzed result indication
In configuration of NM system, E164 is configured as Result Indication is ISDN. Only
coding is adopted for Result Indication replace MCC+MNC with CC+NDC, and leave other
configuration data as they are.
ii. On Translator dialog box, select ISDN/Mobile
numbering plan for configuration of Number Plan.
iii. GT translation data
Direct GT translation data that match
CC+NDC+ H0H1H2H3+XXXXXX to GMSC.
Note: The advantage of the second
configuration method is that the local MSC does
not need to know which HLR it is and needs only
to analyze CC+NDC to GMSC. Then GMSC
analyzes the CC+NDC+H1H2H3 to the destination
HLR GT according to destination CC+NDC+
H1H2H3+XXXXX.

26 Maintenance Experience
www.zte.com.cn

Some Subscribers Couldnt be


Registered in Network after
the Cell Restriction Service Is
added in HLR
Duanmu Xiaoting, ZTE Corporation

Symptom table of VLR was checked, and it was


After cell restriction service is added in HLR, found the value of table R_ZNCD is 100.
some subscribers cannot be registered in the When checking it with the VLR probe,
network while some others can. it was found the number of the current
registered subscribers was 128.
Analysis 4. The number of the cell roaming
1. By checking the HLR, it was found that two subscribers allowed by the local office is
subscribers had the same service subscription defined in the table R_ZNCD, so the extra
information. Therefore, the problem is not caused 28 subscribers could not be registered in
by the subscription information on HLR. the network because of capacity restriction.
2. The problem may have something to do with
SIM card. However, this card could be registered Solution
in the network normally after its number was The carrier demanded that all the
exchanged with another card which is normal. And subscribers should be restricted in one
the other card was then unable to be registered cell, that is, the cell restriction should be
in the network. The registration in the network valid for all the subscribers. So it was
became normal after the subscribed cell restriction required to change the capacity of the
in HLR was removed. table R_ZNCD. After that, when performing
3. It is certain that the fault has something to do the test, subscribers could be normally
with the cell restriction service. Later the capacity registered in the network.

GSM Network Products 27


August 2007 Issue 52 Fault Instances

Failure to Login Remote HLR


Agent
Duanmu Xiaoting, ZTE Corporation

Symptom 2. During the installation of background


The ZTE engineers installed a remote maintenance software of remote client, the area
HLR Client in another city through a router. code and the office ID were respectively set as
After the installation was completed, 21 and 20. Then the IP acquired automatically
basic configuration management could was 192.168.20.210. Meanwhile, 219.192.2.129
be logged in but the agent could not be was selected for IP of 129. After the installation
logged in. The failure as prompted is was completed, the ZTE engineers were able to
failure to connect the server. log in the client and open the interface of basic
configuration management but failed to log in the
Analysis agent with a prompt indicating failure to connect
1. This remote client reaches the to the server. The agent connected to DB through
area where server 129 is located through DBIO, the area code and the office ID set on 129
a satellite. Server 129 connects to the were respectively 888/2 .They are inconsistent with
satellite through a router. The IP of server the area code and the office ID set on the remote
129 is 219.192.2.129. The gateway uses client, which led to the failure in connecting to the
the IP of the router port that is connected server.
with 129, which is 219.192.2.254. The
remote client must use an IP in the same Solution
network section from satellite, which is On this remote agent, the area code and office
192.168.20.210. The gateway uses the IP ID in the file TCPSEEK.INI under C:\ZXG10 were
from the satellite, which is 192.168.20.1. changed to 888/2, and the problem was resolved.

28 Maintenance Experience
www.zte.com.cn

HLR Agent Startup Error When


Configuring Data
Zhang Songlei, ZTE Corporation

Symptom
In the procedure of data configuration
during commissioning, the ZXG10-MSS Agent
Management process of HLR agent kept flickering.
The database can not be linked, and the agent can
not start up.

Analysis
1. It was checked and confirmed that the
HLRDB had been setup.
Figure 1. Error Alert
2. The configuration of HLRDB node which was
managed by HLR configuration was checked and
no configuration data were found.
3. When HLRDB node was being configured,
the system popped up error alert as shown in
Figure 1.
4. Based on the error alert, it was known that
Figure 2. No License Configuration Data
the number of subscribers on the database node
had exceeded the number of subscribers of the
whole system and no matter how small a value the
number of subscribers on the node was changed
to, the system kept giving this alert. It was that the
absence of license might be the reason.
5. Then the license file was loaded and the
system gave an alert as shown in Figure 2 and
Figure 3.
6. It was found that HLR number did not match
that in the License, and then HLR number was
modified. It may be modified in the following ways.
If possible, modify HLR number in HLR
parameter configuration, as shown in Figure 4.
If HLR number can not be modified in HLR Figure 3. Failure to Acquire License

GSM Network Products 29


August 2007 Issue 52 Fault Instances

parameter configuration, directly modify the


parameter in SQL database. Operate as shown
in Figure 5 and Figure 6.

Solution
1. The License could be loaded successfully
after modifying HLR number. The process of
loading License is shown in Figure 7.
2. The database node could be configured after
loading License, as shown in Figure 8.
3. After the configurations above were all done,
the ZXG10-MSS Agent Management process did
Figure 4. Modifying HLR number in Parameter Configuration not flicker anymore. The agent could be opened
normally.

Figure 5. Modifying HLR Number in SQL Database-Step 1

Figure 6. Modifying HLR Number in SQL Database-Step 2

Figure 8. Configuring Database Node Figure7. Loading License Successfully

30 Maintenance Experience
www.zte.com.cn

Failure in Backing up HLRDB


onto Node 129
Chen Yu, ZTE Corporation

Symptom 4. It was found that the file in the


The backup database of one HLR was at node backup directory was much smaller than
129. One day, the system failed to back up the other files and its date was the time when
data automatically to node 129 from node 173. this failure happened the first time.
Checking database showed that the backup MSDB
database file of node 173 saved at node 129 was Solution
suspended. Rebuilding HLRDB and recovering 1. The file containing incorrect data
the data of the last day could solve the problem format was manually deleted.
temporarily. 2. Remote backup was performed
However, one week later, the problem occurred manually after 1:00 am on the day before
again and lasted for several weeks. the next occurrence day of the failure (the
automatic remote backup starts at 1:30).
Analysis The manual backup was successful. Then
1. The backup failure kept repeating, so the automatic backup became normal. And
engineers started from analyzing the error logs of the failure did not occur again.

backup.
2. Check box Show step details at the top right
corner of the log interface, as shown in Figure 1,
was selected. The detailed causes of backup failure
were checked. As the highlighted part showed in
Figure 1, the corresponding data file of node 173
in the backup directory on 129 had incorrect data
format.
3. Through analysis, it was found that the
database format was destroyed in occurrence
of the first failure. Every week, after the problem
happened, HLRDB was rebuilt but the data in
incorrect format was also imported to the HLRDB
backup on node 129 (the data still can be imported
though its format is wrong). Because this file was
never updated by new backup successfully, only
auto-backup was affected. Figure 1. Database Log

GSM Network Products 31


August 2007 Issue 52 Fault Instances

HLR MAP Process Repeatedly


Restarting
Xu Gaofeng, ZTE Corporation

Symptom network cable, disabling and then enabling NIC,


The HLR service process kept all proved useless. Engineers used command
restarting after HLR server power down. ipconfig all in DOS mode and found there was no
IP address. It was doubted that the NIC might had
Analysis a problem.
1. It was doubted that it was caused by 6. Pinging 127.0.0.1 successfully proved that
the exceptional power down, so the engineer the NIC was fine. Changing the network cable
restarted the MAP server. However, the between server and LAN Switch but the problem
process kept restarting after that and the kept the same.
problem remained unsolved no matter how 7. There were network cables of other servers
many times the server was restarted. and their networks were normal, so the LAN
2. Error logs under C:\trace on the SWITCH was fine.
server were checked and only SERVER 8. By checking IP configuration it was found
RESTART alarms were found. that the server had dual-IP (the MAP processor has
3. OMC alarm subsystem was checked two service processes).
and no alarm was found.
4. Re-installation of service handling Solution
programs on the server could not solve the Engineers deleted one of the two IPs and
problem either. waited for the NIC to activate. After one of the
5. Engineers used PING command on service process started, engineers added the IP
the MAP server to check the connection address deleted before and then waited for the
to database server and OMC server but start of the second service process. The problem
failed. Pulling out and then inserting the was solved.

32 Maintenance Experience
www.zte.com.cn

Failure in HLR Setting PDP


Contexts in Batches
Chen Yunlong, ZTE Corporation

Symptom PDP address type (0: dynamic, 1: static)


When configuring PDP contexts in batches IP address (valid when PDP address is
by using a text file on HLR agent, the system static, ignore it when PDP address is
prompted: error occurred when reading dynamic)
subscription file, maybe the format is incorrect. The three parts should be separated
with one or more spaces or tabs.
Analysis
The version of HLR was V3.04.15.P1.B1. In this Solution
version, when a text file is used to perform batch Engineers input the subscriber
operation, the input format will be changed and information in correct format into a text file,
formed by the following three parts: and then successfully set PDP contexts in
IMSI or MSISDN batches.

CUG users Unable to Update


Location under ZTE MSC
Duanmu Xiaoting, ZTE Corporation

Symptom Analysis
A CUG user of an HLR could not update 1. CUG refers to Closed User
location when the user roamed to ZTE MSC1/ Group. Its purpose is to restrict network
VLR. Signaling tracing showed that after HLR communication in the group. Users mainly
sent insert subscriber data to VLR and VLR sent contain multiple CUG members (the
back subscriber data insert Ack, HLR did not send maximum is 10). Normally, the members
Update location Ack. The operator checked the can only communicate with each other
subscription information of this user but found no within this group, but can not communicate
exception. Switching this user to another MSC\VLR with a user out of the group unless the
system also proved useless. user has subscribed for extra functions,

GSM Network Products 33


August 2007 Issue 52 Fault Instances

for example OA and IA. Communication CUG information. So the first package contained
restriction is also available for the the detailed content of CUG SubscriptionList, while
members of a group by subscribing for the second package contained CUG FeatureList.
OCB or ICB. As CUG SubscriptionList is a necessity of CUG
2. Concerning inserting user data, information, the second package contained a
protocol GSM0902 regulates: label of SubscriptionList. ZTE VLR did not support
In any dialogue, the first insert signaling whose subscriptionList was 0 in the
Subscriber Data message which contains second package. This resulted in the failure in
CUG information shall include a non- decoding and the failure in location update.
empty CUG-SubscriptionList. 4. Now it can be concluded that it was a
Th i s i n stru cti o n i n d icates that if compatibility problem on MAP processing in GSM
user data ISD req in the first package protocol. The failure was caused by the different
inserted contains CUG information, CUG- understandings to this protocol by ZTE MSC and
SubscriptionList should not be null. There is Siemens HLR.
no much restriction on the second package.
3. By analyzing the signaling stream, it Solution
was found that the peer end HLR sent two 1. If there are not too many users of this kind, a
ISD req messages, both containing CUG temporary solution could be adopted. First do not
information. That is to say, the detailed provide this user with CUG service or provide this
CUG subscription information existed user with only a part of CUG service, then after
and was inserted at the first package successful location update, add CUG service for
and Siemens HLR inserted CUG label the user. In this way, one message, which contains
in the second package. The point is, all the CUG information, is inserted into VLR by
due to the capacity of one package, the HLR and the location can be updated successfully.
content was divided into two packages 2. This problem will be solved in the next
and the division point was in-between version update by ZTE.

34 Maintenance Experience
www.zte.com.cn

Subscriber Couldnt Access the


Network After HLR Cutover
Duanmu Xiaoting, ZTE Corporation

Symptom length had some errors.


International roaming subscribers could not 7. GT length was checked and it was
access the network after the cutover of a new HLR. found that a GT was 3 digits to 19 digits
And before the cutover, the network was normal. in length. So all the international numbers
of one digit length (numbered from 1 to 9)
Analysis had a problem.
1. Subscription information was checked. It
was found the subscribers had the authority for Solution
international roaming. The GT length was changed to 1 digits
2. According to the roaming area information to 19 digits. After synchronization, no
provided, the engineer checked HLR configuration more SCCP notifications appeared and
and confirmed that country codes allowing roaming the international roaming service became
and access numbers have no problem. normal.

3. Signaling trace on a single subscriber


showed that the process stopped when the
international office requested authentication at
first and then HLR returned three authentication
parameters. The HLR did not receive the location
update request.
4. It was doubted that the authentication
parameter signals were not sent to the switch at
the roaming area. So the engineer asked HSTP
office to check if the HLR successfully sent these
signals to the destination. Later, the engineer knew
HSTP had not received these signals.
5. After analysis, it was doubted that the HLR
had not sent out the authentication parameters
successfully, although they were generated.
6. Alarm observation showed many SCCP
notifications in the alarms and that the roaming
area GT41794908000 tracked was also in the
alarms. Analysis on SCCP notifications showed
that the problems all resulted from the international
numbers. So it was estimated that the setting of GT

GSM Network Products 35


August 2007 Issue 52 Fault Instances

Illegible Codes Appears When


the Network Management
Client Runs HLR Configuration
Management
Duanmu Xiaoting, ZTE Corporation

Symptom language of this computers operating system was


Illegible codes appeared when a client Cyrillic, which was not consistent with the language
was running HLR configuration management. applicable in ZTE software.

Analysis Solution
1. The fault did not disappear after the 1. The default language was changed to West
client software was reinstalled. Europe and United States.
2. REGIONAL OPTIONS were 2. Then the computer was restarted. The
checked and it was found that the default problem was solved.

Unsuccessful Cross Number


Assignment on the two HLR
Databases
Duanmu Xiaoting, ZTE Corporation

Symptom configuration, it was found that the two database


The two database nodes 170 and 175 nodes were in the same number segment, which
could not make cross number assignment was correct. And it was also found that the nodes
successfully. connected to each other successfully. But an
error occurred when the following commands
Analysis were executed on node 170 to update the system
1. By checking the number segment variables on node 175, indicating that the MSDTC

36 Maintenance Experience
www.zte.com.cn

process was not running well. checked and it was found they were in
set xact_abort ON manual status and were not running.
begin tran
update hdb175.hlrdb175.dbo.sysvar set Solution
vinteger=0 where id=1 The problem was solved after
commit tran running the MSDTC process on the two
2. MSDTC processes on the two nodes were computers.

Subscribers Could not Modify


Access Mode in HLR
Su songme, ZTE Corporation

Symptom PDPINFO had subscriber data and the


An office complained that the agent could access mode was GSM only. That is to
not modify the network access mode of some say, subscription data of the subscribers
subscribers. DBIO returned a database execution were not consistent:
failure message when the access mode was select count(*) from baseinfo where
changed from the GSM only to both GSM and nam=1 and imsi in (select imsi from
GPRS. The engineer checked OMMSEE and found gprsinfo)
conflicted primary keys. But the network access select count(*) from baseinfo where
mode of the subscribers whose numbers were re- nam=1 and imsi in (select imsi from
assigned on the local agent could be changed pdpinfo)
successfully. nam=1 denotes querying subscribers
who access GSM only. And 2689 records
Analysis were found, which do not accord with the
1. After accessing the network by dialing up, the data consistency.
engineer queried table GPRSINFO with the given 3. The engineers did not find problems
numbers and found that the subscribers IMSI and of this type any more after modifying the
ISDN already existed in the table. access mode and assigning numbers. So
2. In a normal condition, if the subscribers they estimated that the problems resulted
have just subscribed for GSM network access from the wrong number assignment.
mode, tables GPRSINFO and PDPINFO contain
no records of these subscribers. The engineers Solution
executed the following commands with a query From the queried result, it was
analyzer and found that tables GPRSINFO and necessary to clear up 2689 inconsistent

GSM Network Products 37


August 2007 Issue 52 Fault Instances

d a t a i n t a b l e G P R S I N FO and table 4. Delete the subscribers with inconsistent data


PDPINFO. Before doing so, the local by executing the following sentence (select imsi
database need to be backed up, and then from baseinfo where isdnnum<>0 and nam=1).
the HDB data on node 150 need to be After deleting every time, please check the returned
copied to node 129. The procedure is as results. If some errors presented, note them down
follows: and tell ZTE personals.
1. Back up HDB on node 150 locally 5. Execute the select sentence in step 2
and copy it to node 129. to check if there were inconsistent data in table
2. Select the subscribers needed to GPRSINFO.
be deleted in table gprsinfo (please select 6. Execute the following sentences to check if
HLR subscriber database on node 150). there were inconsistent data in table PDPINFO:
Select count (*) from gprsinfo where select count (*) from pdpinfo where imsi in (select
imsi in (select imsi from baseinfo where imsi from baseinfo where isdnnum<>0 and nam=1).
isdnnum<>0 and nam=1). 7. Delete the rest inconsistent data with the
3. Check if there was a trigger for same method. After clearing up the junk data in
deleting in table GPRSINFO by opening GPRS and modifying the subscriber access mode,
the trigger in the table. the problem was solved.

VPN Subscribers Could not be


the Calling Party When Roaming
Chen Yubin, ZTE Corporation

Symptom resulting in call failure.


VPN subscribers could not be the
calling party when roaming. Solution
Since VPN subscribers are usually VIP users, it
Analysis is recommended to subscribe continue CALL instead
For the VPN subscriber roaming, of release CALL. The advantages are as follows:
HLR will insert CAMEL information during Calling could be continued in case of failure in
location update if the VLR of the called triggering services.
party supports CAMEL2. Then, HLR will When the subscribers make a local call, it can
trigger the intelligent service according also prevent wrong charging by continuing the
to the subscribers OCSI information in CALL instead of answering the CALL when the
VLR. But, for the problems in operator trigger service failed.
network, the signal triggered could not be If SCP is down, allowing continue CALL will not
sent successfully from SSP to SCP, thus influence the VIP services.

38 Maintenance Experience
www.zte.com.cn

Echo Appearing for Outgoing


Calls
Duanmu Xiaoting, ZTE Corporation

Symptom 3. Although the newly added 2M


Echo appeared when outgoing calls were outgoing line was connected to DTEC, a
made. new PP program was loaded again. So
the problem must result from wrong DTEC
Analysis jumpers.
1. Echo appeared when some outgoing calls
were made, but no echo appeared for intra-office Solution:
calls. This symptom indicated that the problem was The board is B0107; the right jumper
associated with CN instead of RNC. should be set at pins 1 and 2. But the
2. Echoes were found just recently. It was actual jumper pins were pins 2 and 3.
doubled that the problem resulted from a 2M Echo disappeared after resetting the
outgoing line which was added recently. jumpers.

Unsuccessful Location Update


for Roaming
Chang Tao, ZTE Corporation

Symptom As expected, by tracking SCCP of HLR,


Mobile subscribers could not access the the engineer confirmed that HLR had sent
network under the control of VLR of other XUDT message and the equipment of the
equipment when they were roaming. It became called party gave no response.
normal after their data had been deleted.
Solution
Analysis 1. The TM PACKET size (under security
The signals were checked and it was found variablesTM variables) was changed to
that there were nine supplementary services in the 140 for the first packet, and the size of the
data inserted, but there were only six after deleting second packet was set to 190.
their data and assigning numbers again. So, it was 2. The HLR did not send XUDT
doubted that HLR sent an XUDT message, but message after the above modification, and
the equipment of the called party did not support. the location update was normal.

GSM Network Products 39


August 2007 Issue 52 Fault Instances

Difficulty in Making Calls at


Busy Hours Under an HLR
Zhou Quan, ZTE Corporation

Symptom might have something to do with database


An HLR office reported that subscribers performances. Database logs were then checked
under this office had difficulty in making and it showed that a large amount of HLRDB170
calls at busy hours, and the office failed to size was full.
modify data through BOSS. 3. The operation failed when trying to apply for
space for tables CAMELINFO and SSDATA, which
Analysis explains the failure in BOSS inserting data.
1. Alarms checking showed alarms of
overhigh UB occupancy and serious traffic Solution
control while actual traffic was just a little The BOSS and the database calls became
heavier and almost as usual. normal when the size for HLRDB170 was enlarged,
2. It was guessed that the problem resulting in normal calls.

HLR System Error Occurred in


Service Subscription on Agent
Yang Guang, ZTE Corporation

Symptom deleted, some other operations also became


During supplementary service testing, invalid, and signaling trace showed no
an HLR system error occurred when information.
operations service subscription, service Two numbers assigned on agent were found
activation through terminal (when by checking R_PCSSSDATA while only one
acknowledgement tone was heard), number was found by checking R_PCSSSINFO
service deactivation through terminal and again two numbers were found by checking
(when acknowledgement tone was physical r_pcsssinfo. All operations became
heard), and service de-subscription normal again after HSDB restart. But the
through agent were performed for operations might be performed twice only and
the third time. The assigned numbers then the HLR system error would appear again if
under test were then unable to be the operations were performed for the third time.

40 Maintenance Experience
www.zte.com.cn

S10SEE showed: CLOSED. The causes of Memory DB


Pno:104, Module:dbiosrv, LEVEL:3, Client not synchronized to physical DB were
type:AGENT, Client node:151, Operator: hereby found.
HLR, ErrCode:150103, ErrInfo:HSDB 2. According to S10SEE print
operate failed(ErrCode:108). memory DB not information it was obvious that the
synchronized to physical DB, Operate failed, physical database was not synchronous
MML: MOD PUSEREX: PSID=481243601118, with the memory database. Since the
DynDataFg=1, ASS=0 foreground memory database reads
the physical database once the HSDB
Analysis restarts, the subscription operations
1. The following operations were performed: may be performed successfully within
Failure observation a short period of time. But after several
Checking capacities of tables in network times of subscription and de-subscription
management system configuration data. operations the memory database changes
Checking configuration of service server can not be synchronized to the physical
node (DB/DBE nodes are configured separately if database, and hence no more subscription
DB_DBE separation is selected). operations may be performed.
Checking DBIO configuration file (checking
HSDB connection also if HSDB module is Solution
configured on foreground). 1. Static route was modified and actual
Checking DBE configuration file (similar to office ID was added to the monitor script,
checking DBIO). as shown below:
Checking connection between DBE and DEST=130.0.0.0
foreground and connection between DBE and NETMASK=255.0.0.0
DBIO by running S10SEE on the micro computer. GATEWAY=129.0.2(Office ID).100
It was finally found that DBE did not 2. The problem was solved and
communicate with the foreground and that S10SEE the DBE communicated well with the
showed all connections with the foreground were foreground.

GSM Network Products 41


August 2007 Issue 52 FAQ

How to Analyze the Location


Update Faults
Wang Dankai, ZTE Corporation

A: If only concerning the faults occurring is found that the opposite-end does not return
at the switching side, analyze the fault the signaling, without configuring the data of the
from the following aspects: local end or returning the authentication response
1. The link at the A interface is faulty, message;
resulting from the transmission or the 4. Due to the problem of version compatibility, the
abnormal working status of the board; opposite end does process or discard the location-
2. No number analysis is performed on update-related signaling sent from the local end;
the IMSI; 5 . R e s t r i c t e d b y t h e V L R c a p a c i t y, t h e
3. Through the signaling tracing, it subscriber location update fails.

How to Handle Unclear Blocked


Signaling Links
Wang Dankai, ZTE Corporation

A: If the signaling link is unclear, the system impedance matching of the PCM
engineer can handle it in the following ways: system where the signaling link is located is
Check whether there is any alarm in correct.
the transmission system of the PCM Check whether the local MTP data are correct
system where the signaling link is and whether the signaling route and the
located. signaling office direction are configured.
Check whether the link IDs, timeslots, Check whether the links of the local office and
and transmission system IDs of the the opposite-end office are blocked.
local office and the opposite-end office Check whether inter-module communication is
are consistent. interrupted.
Check whether the transmission Check the data of the opposite-end office.

42 Maintenance Experience
www.zte.com.cn

How to Set the Dual Seizure


Handling Mode
Wang Dankai, ZTE Corporation

A: The handling mode of the dual seizure should be seizure handling mode 1. In this case, the
negotiated by both sides. The dual seizure handling office with small SPC seizes the circuit in
mode 1 and the dual seizure handling mode 2 are the ascending order in each PCM, and the
provided in the indicator of the trunk management. office with big SPC seizes the circuit in the
If the parity control is adopted by negotiation, select decreasing order in each PCM.
the dual seizure handling mode 2. At this time, the After the handling mode of dual seizure
office with small SPC seizes the odd circuit first, and is understood, the specific option is
the office with big SPC seizes the even circuit first. selected from the Select Method of Trunk
If the sequence control is adopted, select the dual in the trunk configuration.

How to Transfer the Database


Wang Dankai, ZTE Corporation

A: Due to some reasons, it is required to transfer the database to other directory. The path of the
database equipment is stored in the sysdevices table of the master database. Two lines in the
sysdevices table are shown as follows.

low high status cntrltype name phyname mirrorname stripeset


721420288 721440767 2 0 aaa D:\MSSQL\DATA\aaa.dat
721444767 721464590 2 0 aaalog D:\MSSQL\DATA\aaalog.dat

Therefore, to transfer the database, just modify the phyname in this table.
For example, to transfer the above-mentioned aaa database from D:\MSSQL\DATA to E:\
DATA, execute:
use master
go
sp_configure 'allow updates', 1 Allowing modifying the data in the system database
go

GSM Network Products 43


August 2007 Issue 52 FAQ

RECONFIGURE WITH OVERRIDE Configuration takes effect


go
UPDATE sysdevices set phyname = 'E:\DATA\aaa.dat' WHERE name = 'aaa'
UPDATE sysdevices set phyname = 'E:\DATA\aaalog.dat' WHERE name = 'aaalog'
Go If the log and DB exist on the same equipment, only one piece of equipment is needed to
be updated.
sp_configure 'allow updates', 0 Forbiding modifying the data in the system database
go
RECONFIGURE WITH OVERRIDE
Go
Stop the SQL server, copy the file aaa.dat and the file aaalog.dat to the new directory, and then
restart the SQL server.

How to Recover the Database in


the Suspect Status
Wang Dankai, ZTE Corporation

A: If the database is in the suspect status, recover it first through the following way:
use master
go
sp_configure 'allow updates',1
go
reconfigure with override
go
update master..sysdatabases set status = status ^ 256
WHERE name = dbname
go
sp_configure 'allow updates',0
go
reconfigure with override
go
Restart the SQL. If the database still cannot be recovered, the database is already damaged,
or the database file does not exist. Delete the database first, and then recreate it, and recover it
from the backup file.

44 Maintenance Experience
www.zte.com.cn

GSM Network Products 45

Vous aimerez peut-être aussi