Documente Academic
Documente Profesional
Documente Cultură
Fault Instance
Some Pre-Paid Subscribers Could Not Receive Calls 2
Active/Standby Servers of the Database Could Not Connect to Disk Array Simultaneously 3
Trunks between MSCs Were Busy 4
Abnormal SMS of an International Roaming Subscriber 6
Full R_CUG Table Caused Location Update Failure 7
Data Backup Operation on MSC Client Failed 7
Charging Server Switched Off Automatically 8
New Added 156 Number Section Could Not Attach GPRS Network 9
SMS Interconnection Failure 10
SDTB Board Alarm 11
MSC CDR was Lost 11
Experience Exchange
MSC Sent Overload Alarm to BSC 27
March 2008 Issue 89 Fault Instance
Maintenance Experience
www.zte.com.cn
of the disk array was short, so that the array, there are two fiber interface modules. Each
active/standby server could not connect fiber interface module provides two servers with
to the disk array at the same time. After two sockets. In general, only one fiber module
switching the active/standby fiber module on the EMC is configured, and the other module
connecting to the back of the disk array, is in the idle state. Before leaving the factory, the
the problem was solved. configuration is that only one fiber module can
Note: At the back of the EMC disk connect to the server.
Symptom so that the call from the MSC-A to the MSC-B was
The trunk between MSC-A and rejected by the MSC-B.
MSC-B was busy, which affected the call According to the above judgment, it was
completion ratio. deduced that there was something wrong with
the MSC-B. Maybe the trunk between the MSC-B
Analysis and other MSCs was busy. When analyzing the
1. After checking the data configuration performance data for each trunk group of the
and the security variables between MSC-B, the engineer found that the busy ratio of
these two MSCs, and comparing these interoffice trunk among all the trunk groups (except
configurations with other MSCs, the MSC-A (IN)) was normal.
engineer did not find any problems. i. Open the failure observation when the trunk
2. When checking and analyzing the is busy and calculate the call loss data for
performance data, the engineer found that 5 minutes. On the MSC-A, calculate the
the busy ratio of the trunk from the MSC-A performance data for five minutes and then
outgoing to MSC-B incoming was very calculate the busy times of interoffice trunk to
high. However, the traffic between these MSC-B.
two MSCs was not heavy. It was doubted ii. Extract the failures (2034=No circuit work
that there was something wrong with the (NoCircuitWork)) from all the call loss data, as
trunk from the MSC-B to another MSC, shown in Figure 1.
Maintenance Experience
www.zte.com.cn
Summary
In the CDR record, not all the CDR will be
charged, so it is required to check the detailed
fields in the CDR.
Maintenance Experience
www.zte.com.cn
Analysis Solution
When analyzing the call loss, the engineer The problem was solved after the
found a lot of 15=Cug Rejected records. After engineer expanded the capacity of the
checking the capacity usage of each table in table related to CUG.
Symptom
When backing up the basic configuration data
with the BCP method on the client, the warning
unable to open the file appears, as shown in
Figure 1. After clicking the OK, the prompt backup
fail,SQL communication may be interrupted or the
disc id full appears, as shown in Figure 2.
Analysis
1. On the 129 server, the data backup with the
BCP and SQL methods was normal. The above Figure 1. Warning
existed.
3. After checking the data configuration related
to BDE, the engineer did not find any abnormalities.
4. When checking the update about the anti-
virus software, the engineer found that the anti-
Figure 2. Info
virus software had not been updated for three
months. It was doubted that the virus caused the
error prompt meant that the SQL database SQL database failed to connect to the server.
failed to connect to the server. However,
the data backup with the method SQL Solution
was normal, which meant that there was When updating the virus database on the client
something wrong with the client instead of and scanning the disk, the engineer found some
the program of the server. worm virus infections. After deleting this virus and
2. After re-installing the version re-starting the client, the engineer found that the
program of the client on the client, the data backup with the method BCP was normal, that
engineer found that the problem still is, the problem was solved.
Maintenance Experience
www.zte.com.cn
Solution
After checking and analyzing the configuration,
the engineer found that the option SC
Support under Basic Configuration Local
Configuration Mobile Data Configuration
MORE of MSC was unchecked. After selecting
this option and synchronizing the data, the SMS
function was normal.
10 Maintenance Experience
www.zte.com.cn
wireless side, the alarm disappears. J1 Char [192] High rank channel trace information
Networking
The charging system of the current network
adopts the call meter mode. The corresponding
principle figure is shown in Figure 1.
The foreground generates the CDRs according
to the following three methods:
i. Back up the CDR to the directory ZXGMPBAK Figure 1. Billing System
of the charging server in the form of did not find the CDRs between May 31st and June
the original CDR. The extension name 22nd.
of the file is .zzz.
ii. Back up the coding/decoding GCDR Analysis
file to the directory ZXGBILL. When checking the ZXGMPBAK directory
iii. Output the foreground CDR to the on the site, the engineer did not find the zzz file
directory SRCBILL through the billout between May 31st and June 22nd. However, under
program. The extension name of the this directory, the size of the zzz file generated on
file under this directory is .jb. June 23rd was bigger than the general size. It was
doubted that the accumulated original CDRs about
Symptom these twenty-three days were saved to the CDR
Under the directory M: \zxgbil generated on June 23rd after both the foreground
o f t h e M S C c h a r g i n g s e r v e r, t h e and the background recovered normal.
maintainer found that there was a CDR After copying the zzz CDR file generated on
file UN200705301039372754GCDR June 23rd to the directory SRCBILL, automatically
generated at 11:48, on May 30th. And the extracting it into the jb file, and distilling ZXJ10,
following file was UN2007062317162521 the engineer found the CDRs beginning from May
05GCDR, which meant that about twenty- 31st, which meant that the CDRs were not lost.
two days of CDR files were lost.
After connecting the charging server of Solution
ZXJ10 to the MSC system, the maintainer After copying the original CDR zzz file under
found that the original CDR generated on the directory ZXGMPBAK, automatically extracting
May 30th was 1.112M, and the original it into the jb file, and distilling the ZXJ10, the billing
CDR generated on June 23rd was 3.313M. center completed the settlement and the problem
After the settlement, the maintainer still was solved.
12 Maintenance Experience
www.zte.com.cn
found many HLR session is overtime errors. side cooperated with the maintainer on
According to the performance statistics data of the MSC side and activated the four
SMSC, the engineer found that the HLR session new added signaling links. After that, the
is overtime error was eight times more than before signaling from the HLR was sent to the
from July 10. new module of the short message through
After asking the maintainers on the short the link. At that time, the communication
message side, the engineer found that there was only between the new module and the old
one foreground module before, and a new module module was disabled, which caused that
was added recently. The new added module was the routing request from SMSC to HLR
not used, and was also not connected to the service failed and the failure times multiplied.
processor. However, there was some signaling links
configured between the new added module and the Solution
HLR (four signaling links were added). After blocking the four new added
On July 9, the maintainer on the short message signaling links, the problem was solved.
Symptom
For an office with a single MSC, sometimes
the subscribers under the single MSC could get
through the international calls, but sometimes the
subscribers could not get through the international
calls.
Analysis
When the local subscriber dials the international
subscribers, it is required to add the long-distance
access code (2, 5, 7, and 9) before the international
number. However, when adding these four
prefixes, the problem appears. These four prefixes
are configured in the international analyzer entry 4
of the MSC. Except the outgoing routing link, the
configurations of these four prefixes are just the
same. The outgoing routing links are as follows:
1. 2 out route chain 1 our route set 1 Figure 1. Number Analysis Failure
2. 5 out route chain 1 our route set The local maintainer reported that MSC could
1, 2 not send the IAM message if the call failed, but
3. 7 out route chain 1 our route set MSC could send the IAM message if the call was
1, 2 successful. For the call loss, it was known that the
4. 9 out route chain 1 our route set number analysis failed, as shown in Figure 1.
1, 2 The engineer could pre-analyze the called
When checking the data configuration number through the Preview Digit Analysis
of the MSC (including the trunk group tool in the dynamic management, and then the
configuration and the trunk indicator engineer could find the correct outgoing routing
configuration), the engineer did not found link, as shown in Figure 2.Through the analysis,
any abnormalities. It was doubted that the the engineer found that the call failed if the called
international gateway exchange failed to number was 002160XXXXXXX. As shown in
analyze the called number. Figure 2, the value of Toll Prefix+Area Length is 5,
that is, the toll prefix and the area length is 5. The
following numbers are the SN. However, the sixth
bit of the called number 002160XXXXXXX is 0,
which causes the number analysis failure.
Solution
After modifying the value of Toll Prefix+Area
Length as 3, the engineer found that the problem
Figure 2. Digit Analysis Preview was solved when performing the on-site test.
14 Maintenance Experience
www.zte.com.cn
16 Maintenance Experience
www.zte.com.cn
disconnected physical interconnection: the interconnection of the signaling link was normal.
E1 connection is wrong or the physical data
configuration of the DTI board is wrong. Summary
After eliminating the first reason, 1. When interconnecting with the equipment
the engineer checked the physical data of other manufacturers, it is not only required to
configuration about the interface board check whether the option CRC Verification of
with the peer end again. After coordinating the physical configuration is consistent, but also
with the peer end office and changing the required to select NO CRC (only for ESF) for this
option of the physical configuration CRC option in order to match it.
Verification from CRC (only for ESF) to 2. After the physical connection, it is not only
NO CRC (only for ESF), the engineer required to check whether the E1 indicator on
found that the E1 running indicator flashed the DTI board flashes, but also required to check
frequently and run normally. And the whether the flash frequency is equal.
18 Maintenance Experience
www.zte.com.cn
6. When checking the watchdog log and the deleting some alarms and prompts manually
system log on the 129 server, the engineer did not and expanding another 300 M space for the
find any abnormalities. However, there were many alarm database, the engineer found that this
SQL errors in the application program log of the alarm disappeared immediately.
129 server.
7. When checking the SQL log, the engineer Summary
found many Data table cannot be written During the daily maintenance, it is
error logs. It was doubted that the table space required to check the databases of which
of some SQL data was full, which caused the the space is lack or low. When the data
communication between the foreground and the cannot be written into the database
background interrupted. because of low space, the communication
8. When checking all the SQL databases, the abnormal alarm of the 129 server and the
engineer found that the available space was low. After foreground will generate.
2. Trace the signaling and check the call loss. 2007-10-2 19:00 45.96 95.4
corresponding called numbers from the that the location area was very concentrated,
signaling tracing. but the cell ID was very dispersive, and there
Many subscribers send the short was no regularity.
message to this number (this number iv. A f t e r s o r t i n g 5 5 s u b s c r i b e r s f r o m 3 0 4
NDC 339 does not belong to any carrier), subscribers in the next day, the engineer
so the success ratio of MO declined. It is found that one subscriber was switched on,
doubted that this is an illegal behavior of the information of four subscribes was deleted
the subscribers. from the VLR, the LAC of three subscribes
3. According to the CDR of the current were changed, and the GCI of twenty-two
two days in the short message center, sort subscribers were changed.
all the calling numbers (456 in total) that 4. Through cracking the content of the short
send the short message to the subscriber message, it is doubted that there is some evil
33960XXXXX, and then query and analyze virus in the handset terminal. After the analysis,
the calling numbers. the engineer found that this was really a kind
i. According to the number section of handset virus. The subscriber sent the short
analysis, the engineer found that 305 message to the number 9233960XXXXX every five
numbers (67%) belong to the area-A. minutes when they knew nothing about the virus.
And the engineer knew that there were 5. The subscriber needs to delete the virus from
two BSCs in the area A : B92 and the handset terminal.
B91, and a new BSC B90 was added i. Install the file manager on the handset.
recently. ii. Find the directory \system\apps\commwarrior
ii. When querying these numbers in the from the handset and then delete the files
VLR of area-A, the engineer found commwarrior.exe and commrec.md.
the related information of these 304 iii. Enter the directory \system\updates\commwarrior
subscribers (67%). and then delete the files commwarrior.exe,
iii. After recording the location information commrec.mdl and commw.sis.
of the numbers that can be found from iv. Enter the directory \system\recogs and then
the VLR of area-A, the engineer found delete the file \system\recogs.
20 Maintenance Experience
www.zte.com.cn
This duration can be adjusted through This version could be used to adjust the duration of
modifying the 15# timer and the 162# the paging. At that time, the timer was 169.
timer. After updating the version, it is required to
Total duration of the paging response modify the above three timers (15#, 162# and
waiting = min (T15, T162), which means 169#) and then perform the corresponding test.
that the total duration is the minimum
value of these two timers. It is required to Solution
modify the duration of these two timers as 1. When adjusting the timer, it is required to
13 seconds according to the requirement adjust the timers of all the modules. After the
of the carrier. adjustment, it is required to check whether this
At that time, the on-site version was adjustment operation takes effect.
MSC V3.04.12.P11.B1a. The fixed duration 2. It is not recommended to set a long duration
of the 1st paging and the 2nd paging for the Waiting paging response duration timer.
was 3 seconds. The carrier required to For example, if it is set as 15 seconds, maybe the
set the duration of the 1st paging as 6 MSC could not play tone to the calling party. A long
seconds, so it was required to update the paging duration is also a waste of the resource, so
foreground version to V3.04.12.P11.B1c. the duration should be within 13 seconds.
22 Maintenance Experience
www.zte.com.cn
Symptom Analysis
At present, the subscriber under the When checking the number analysis
local MSC dials the long-distance mobile configuration of the MSC, the engineer found that
subscriber or the PSTN subscriber with both the pre-analysis selector and the MO analysis
the following format 0+ (3, 5, 7 or 9) + selector had used the DAS101. In the DAS101,
area code + number. When dialing the the digit diversion (calling) operation was not
local subscriber, it is only required to dial done for the prefixes 03, 05, 07 and 09. All these
seven digits directly, such as 2XXXXXX, indicated that the number format sent by CC to
3XXXXXX, 4XXXXXXX and 8XXXXXX. VLRMAP was just the 03XXXXX format dialed by
The carrier should subscribe the the subscriber. Furthermore, VLRMAP has deleted
long-distance restriction service for the the digit 0, so it was not matched with the number
subscriber. On the site, the OSB method in the list of OSB service restriction.
was adopted. On the HLR, it was required According to the normal requirement of OSB
to subscriber the OSB service for the service, the format sent to VLRMAP by CC should
subscriber. Under the Mobile Number be CC+ area code + number. MSC needed to dial
Analysis Configuration > ODB and OSB these prefixes, so the following schemes were
Service Call Restriction Number of MSC, given:
the four long-distance prefixes 03, 05, 07 1. Under the ODB and OSB service
and 09 are configured. In fact, the long- configuration of MSC, restrict the prefixes 03, 05,
distance calls were not restricted, but the 07 and 09.
local call of which the prefixes are 3, 5 and 2. In the pre-analysis, add an analyzer entry
7 are restricted. and then analyze the numbers of which the
24 Maintenance Experience
www.zte.com.cn
Solution
1. Under the Configuration Management >
Supported Service Option Configuration of HLR,
check the checkbox Support OSB Service, as shown
in Figure 1.
2. Under the Configuration Management
> PLMN Specific ODB Configuration of HLR,
configure the self-defined OSB service, as shown
in Figure 2.
3. Subscribe the OSB service for the subscriber
in the HLR agent, as shown in Figure 3. Figure 3. ZXG10-MSS Agent System
26 Maintenance Experience
Experience Exchange www.zte.com.cn
On the site, MSC sent the Overload alarm to The information related to the data
BSC for many times. The reasons of the alarm are area is not shown in the performance
as follows: the CPU is overloaded, the value of UB s t a t i s t i c s . H o w e v e r, y o u c a n
increases in an instant, the occupation rate of the dynamically check the occupancy rate
data area is very high or the data area is overload. of the data area of each processing
1. When MSC sends Overload to BSC, layer (such as MM, MAP and CC)
it follows some rules. The flow control has a according to the service observation.
relationship with the CPU, the UB and the data ii. Set the flow control level of UB, as
area. The level setting of each kind of flow control shown in Figure 1.
is shown in Table 1. In general, the usage rate of the
2. Set the flow control in the MSC Security memory is not very high, but the value
Variable.
i. Set the level of the CPU and the data area. Table 1. Flow Control Index
The high CPU occupancy rate means that Flow Control\ Level 6 5 4 3 2 1
the events processing frequency is very high. CPU Load control 60 64 68 72 76 80
When the occupancy rate of the CPU reaches UB max rate control 60 64 68 72 76 100
100%, the MP will be in an endless-loop status, Used Index control - - 2670 2730 2790 2850
28 Maintenance Experience
March 2008 Issue 89 Fault Instance
30 Maintenance Experience