USCDB Principle with

Operation and
Maintenance

www.huawei.com

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved.

 Foreword
l  The Unified Subscriber Center Database (USCDB) consists three
subsystems, including the data service (DS), date storage (also called
DBMS, in short of database management system), and subscriber data
management. PGW (provisioning gateway) is most important of the
subscriber data management subsystem.

l  The DS is the data service module. It implements distributed storage of the

subscriber data and provides reliable data access.

l  The DBMS is the permanent storage module. It stores global data,

subscriber data, routing data, and PGW operation logs.

l  The PGW is the subscriber data management module. It provides subscriber

data management and data configuration of MAP services.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page2
 Objectives
Upon completion of this course, you will be able to understand:

p  Logical structure and basic working principles of the DS(data

service subsystem)

p  Logical structure and basic working principles of the

DBMS(date storage subsystem)

p  Logical structure and basic working principles of the subscriber

data management subsystem

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page3
 Contents
1. Overview of the USCDB
2. DS Principles and O&M

3. DBMS (date storage subsystem) Principles and O&M

4. PGW Principles and O&M

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page4
SingleSDB Software Structure

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page5
Overview of the DS-Interface
Introduction

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page6
 Overview of the DS
Node: A unit that can independently provide services. A host may have one or more
l 

nodes. Based on the node status, the nodes can be classified into two types:
p  Master node: Each cluster can have only one master node. The master node
implements functions such as adding, deleting, modifying, and querying data.
p  Slave node: Each cluster can have multiple slave nodes. The slave nodes can
implement only data querying.
l  Cluster: Nodes manage the same subscriber data form a cluster.
p  A cluster generally consists of one master node and several slave nodes.
p  When the master node is faulty, services are switched over to other slave nodes in
the cluster for processing. Therefore, service processing is not adversely affected
by the failure of a single node, improving the system reliability.
l  NE: A network entity that is independently deployed on other network equipment or
upper-layer network management system (NMS). Both the USCDB and FE function as an
independent NE.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page7
 Overview of the DS (Cont.)
Data services: The DS provides data services, such as adding, deleting, modifying, and
l 

querying subscriber data based on the service processing requirements of the FE. In addition,
it supports transaction-based or non-transaction-based data access.

Data loading: After a DRU or DSU process starts, it creates an in-memory database, loads
l 

data, and then provides services. The DRU or DSU process loads data from the master node,
the local dump files and redolog files, and the physical database in sequence.

Data replication: To ensure data consistency among nodes in a cluster, the master node
l 

sends a replication message to the slave nodes after modifying the data. The slave nodes
then update the data in the local in-memory database according to the replication message.
Data synchronization: To ensure data consistency between the in-memory database and
l 

the physical database, the master node records a synchronization log after modifying the data
in the in-memory database and the synchronization module modifies the data in the physical
database according to the synchronization log.
Redundancy function: The USCDB supports the active/standby redundancy solution and
l 

load-sharing redundancy solution.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page8
DBMS Interface Introduction

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page9
Overview of the DBMS
Physical database
l 

The physical database is a device for permanent data storage. It is

p 

independent from service processing. The physical database stores

the global data, subscriber data, and routing data, and PGW
operation logs. It uses the PT(Proton) database management
software and stores data on the disks of VMs or the disk array.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page10
 Overview of the DBMS(Cont.)
Functions of modules in the physical database
The physical database consists of the DBG module, NDF module, Oracle
l 

database, and disk array.

p  DBG module
Serves as the access point to connect to all the DRU and DSU nodes.
l 

Converts the format of the messages between the physical database and the DRU
l 

or DSU nodes.
p  NDF module
Manages the physical database
l 

Monitor platform resources

l 

p  PROTON: serving as the database management system (DBMS), backs up the

data stored in other modules.
p  Disk array: It uses 300 GB hard disk as the storage medium and adopts the
RAID10+Hot spare solution. Twelve hard disks can provide a maximum of 1500
GB storage space.
Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page11
PGW Interface Introduction

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page12
 Overview of the PGW
PGW Overview
l 

PGW is the most important module of the subscriber data management subsystem,
which serves as the BE access interface. It performs the following functions:
p  Converts formats of the operation commands received from the provisioning
system into the format required by the convergent database based on the user data
model.
p  Verifies the validity of data based on the predefined subscriber data.
p  Authenticates and authorizes the provisioning system.

Function of modules in the PGW subsystem

l 

The PGW subsystem consists of the DPU module, and the PGW module.
p  DPU module
p  PGW module

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page13
 Contents
1. Overview of the USCDB
2. DS Principles and O&M

3. DBMS (date storage subsystem) Principles and O&M

4. PGW Principles and O&M

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page14
 Contents
2. DS Principles and O&M
2.1 Data Replication

2.2 Data Verification

2.3 Data Loading

2.4 Data Backup

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page15
 Data Replication Overview
l  Data replication is a process of replicating the data updated in the local
node to other nodes. This ensures data consistency between the nodes
in the same DRU or DSU cluster and between the active and
redundancy USCDBs.

l  Data replication can be classified in the following ways:

p  Based on the replication mode, data replication is classified into

synchronous replication and asynchronous replication.

p  Based on whether the data needs to be synchronized between
locations, data replication is classified into inter-location replication
and intra-location replication.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page16
Data Replication Principle
l  Synchronous replication
p  Synchronous replication is the process in which the master node
sends a replication message to the slave nodes and waits for the
responses from these nodes. On receiving the responses from these
nodes, the master node continues to process the next replication
request.

l  Asynchronous replication
p  Asynchronous replication is the process in which the master node
writes replication messages into the asynchronous replication cache
queue and proceeds with the subsequent processing without waiting
for the asynchronous replication result. The management thread of the
cache queue periodically scans and sends replication messages
Copyrightretained in the
© 2017 Huawei cache.
Technologies Co., Ltd. All rights reserved. Page17
 Data Replication Principle (Cont.)
Master 1’
2’ Slave
1 2
4
Asynchronous 3 <=Loading
replication cache

Slave

Asynchronous replication procedure

l 
Synchronous replication procedure
l 
Asynchronous replication is performed when the peer
Synchronous replication is performed when the peer
node is not in Slave or Premaster state.
node is in Slave state.
The master node writes the replication message into
l 
The master node sends a synchronous replication
l 
the asynchronous replication cache.
message to the peer node.
The master node sends an asynchronous replication
l 
After replicating data from the master node, the
l 
message to the slave node.
slave node returns a response to the master node.
After replicating data from the master node, the slave
l 

node returns a response to the master node.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page18
 Data Replication Principle (Cont.)
l  Intra-location replication
p  Intra-location replication is the master node replicates data to the slave nodes in
the same cluster. In the same cluster, only the master node can update data. When
updating the data, the master node needs to replicate the updated data to the slave
nodes in the same cluster.

p  In a location, static data adopts the synchronous replication mode, and dynamic
data adopts the asynchronous replication mode.

l  Inter-location replication
p  Inter-location replication is the active USCDB replicates data to the redundancy
USCDB. After the data is updated on the master node of the active USCDB, the
updated data is replicated to the corresponding master node with the same cluster
ID in the redundancy USCDB.

p  Data replication between locations adopts the asynchronous replication mode.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page19
Data Replication Principle (Cont.)
l Inter-location replication procedure

Global data
replication of
DRU cluster

Subscriber
data replication
of DSU cluster

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page20
 Data Replication Principle (Cont.)
Collecting information about replication messages
l  1. Open the Performance Management System page, and choose USCDB[ME] >
USCDB Measurement > USCDB DCI Operation Measurement, select a master node
and view the number of Replication Positive Requests Sent, select the related slave
node and view the number of Replication Positive Requests Received and the ratio of
Success Ack Ratio of Local Node Replication.

l  2. Run DSP NODE and query the value of Size of verify queue.

l  3. View Counts of Asynchronic Replication Message in the service plane.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page21
Operations Related to Data
Replication
l  Scenarios

p  Data replication is performed to ensure data consistency between the

active and standby DRU/DSU nodes and the active and redundancy
USCDBs.

l  Prerequisites

p  None.

l  Procedure

p  No manual operation is required because data replication is

performed automatically.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page22
 Contents
2. DS Principles and O&M
2.1 Data Replication

2.2 Data Verification

2.3 Data Loading

2.4 Data Backup

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page23
Data Verification Principle
l  Overview of Data Verification
p  Data verification is performed to ensure data consistency between
master and slave nodes, DRU/DSU nodes in the same cluster and data
consistency of global data between the FE and BE. Data verification is
classified into the following two types:

p  GU-HLR global data verification

p  USCDB data verification: data consistency check between and within

DRU/DSU nodes in the same cluster, it is also classified into four types:
n  Exception check
n  Periodic data check
n  Single subscriber check
n  Data consistency check between the in-memory and physical databases

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page24
 Data Verification Principle –GU-HLR
Global Data
l  Principle
l  Global data verification is performed when the global data between the FE
and BE is inconsistent.
p  When the CCU processes are started, a specified CCU process (known as the
loading CCU) loads the global data from the DRU node and stores the data in
the memory.

p  Other CCUs (known as common CCUs) load the global data from the loading
CCU.

l  The loading CCU updates the global data from the DRU and common
CCUs updates global data from the loading CCU every 20 seconds. If a
common CCU fails to load data from the loading CCU three times, the
common CCU directly loads data from the DRU.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page25
 Data Verification Principle -Exception
check
l  During data replication, if the master node fails to receive a response from the
slave node within the specified time or fails the replication, the master node
records the exception check logs. Then, the master node initiates an exception
check.
l  Exception check Procedure
3
Master Slave

1 2

Message
cache

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page26
 Data Verification Principle -Periodic
data check
l  Periodic data check is performed to ensure data consistency between the
master and slave node of the DRU/DSU cluster, it is classified into two types:
p  Data check in Single site

p  Data check between the active and redundancy USCDBs (intra-location data check )

l  Data check in single site is classified into the following four types:
p  Horizontal forward check (Initialed by the master node)

p  Horizontal backward check (Initialed by the slave node)

p  Vertical forward check (Initialed by the master DSU node)

p  Vertical backward check (Initialed by the master DRU node)

l  The data check in single site has the following rules:

p  When horizontal check is performed, data is updated based on master node.

p  When vertical check is performed, data is updated based on the master DSU node.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page27
 Data Verification Principle -Periodic
data check (cont.)
l  Inter-location data check between the active and redundancy USCDBs is
performed to ensure data consistency between the USCDBs, it is classified
into the following three types:
p  If a data record is available in the active USCDB but unavailable in the
redundancy USCDB, the active USCDB initiates a horizontal forward check. The
data in the redundancy USCDB is updated based on the data in the active
USCDB.
p  If a data record is available in the redundancy USCDB but unavailable in the
active USCDB, the redundancy USCDB initiates a horizontal backward check.
The data in the redundancy USCDB is updated based on the data in the active
USCDB.

p  If a data record is found inconsistent between the active and redundancy

USCDBs, the data record with a later timestamp will be synchronized.
Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page28
Operation Related to GU-HLR Global
Data Verification
l  Scenarios
p  GU-HLR global data verification is performed automatically.

l  Prerequisites
p  Data verification parameters are configured.

l  Procedure
p  Log in to the PGW Web LMT, and run SET MAPSERV to set
GLOBALTABLEREFRESHTIME.

History Command:
Command Input: Assit Exec

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page29
 Operations Related to Periodic data
check
l  Scenarios
p  DRU/DSU data verification is performed automatically.

l  Prerequisites
p  Data verification parameters are configured.

l  Procedure
p  Log in to HUAWEI Operation & Maintenance System, open the MML
Command - USCDB window, and run MOD INSP to enable or disable data
verification for a FE and set data verification interval.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page30
 Operation Related to Data
Verification-Enable Data Verification
l  Enabling data verification

l  Scenarios
p  To ensure data consistency, another periodic data check can be enabled by
setting related parameters if a periodic data check is not complete.

l  Procedure
p  On the OMU client, run MOD INSP to change the value of VFYCYCPRD to
60s so
MOD thatINSPT=FE,
INSP: another periodic data check will
INSPN=VFYCYCPRD, be performed
INSPV=60, after 60s.
FETYPEN="GU-HLR";

p  After another periodic data check is performed, modify the value of

VFYCYCPRD to the default value.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page31
 Operation Related to Data
Verification-Disable Data Verification
l  Disabling data verification Scenarios
p  Data verification must be disabled before important operations such as migration,
upgrade, and expansion so that data in the active USCDB will not be affected.

l  Disabling Procedure
p  On the LMT, run MOD INSP to disable intra-location and inter-location data verification.

p  Run the following commands to disable the intra-location data verification.

MOD INSP: INSPT=FE, INSPN= FWDVFYSWT, INSPV=0, FETYPEN="GU-HLR"; (Forward verification
switch )

MOD INSP: INSPT=FE, INSPN= REVVFYSWT, INSPV=0, FETYPEN="GU-HLR"; (Reverse verification

switch )
MOD INSP: INSPT=FE, INSPN= FWDRTVFYSWT, INSPV=0, FETYPEN="GU-HLR"; (Forward route
verification switch )

MOD INSP: INSPT=FE, INSPN= REVRTVFYSWT, INSPV=0, FETYPEN="GU-HLR"; (Reverse route

verification switch )
Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page32
p  Run the following command to disable the inter-location data verification.
Operation Related to Data
Verification- Data Verification results
l  Querying check results
p  Log in to the OMU Client, and run DSP MDBVERIFY to query the
result of the data verification on the in-memory databases (IMDBs).

p  Run DSP MDBVERIFY and check the value of Inconsistent Objects,

which indicates the number of inconsistency.

p  Data verification is performed based on the data in the master node in

case of inconsistency.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page33
Supplementary Operations of Data
Verification - performance
measurement tasks
l  Querying data verification performance measurement tasks
p  Open the Performance Management System page, and view
performance measurement tasks related to data verification.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page34
Solution 1 Related to Data Verification
Symptom 1:

l  The DS database is abnormal. All or some DRU and DSU nodes are
red on the USCDB service panel.
Solution: To solve this problem, perform the following steps:

p  1. Check whether flow control for BE is enabled. If the flow control is

enabled, check whether the database becomes normal after the flow control
is complete.

p  2. Disable the data verification and backup function because data

verification and backup have an adverse impact on the performance, which
may cause the abnormality of the database.

p  3. Check whether a batch operation is being performed.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page35
Solution 1 Related to Data
Verification（Cont.）
Symptom 2:

l  The active and redundancy USCDB of a site works in load-sharing

mode. A subscriber is defined and provided services on the active
USCDB, however data of this subscriber is not updated to the
redundancy USCDB in time because the bearer network has poor
quality. When the subscriber access the redundancy USCDB, the
subscribed services cannot be used.
Solution:

l  Run CHK CONSY on the active USCDB to perform consistency check

on the data of the subscriber. The slave node updates the subscriber
data based on the data in the master node in case of inconsistency.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page36
 Solution 2 Related to Data Verification
l  Symptom: Data is updated on the active USCDB but is not synchronized to
the redundancy USCDB.

l  Solution:
p  1. Open the Browse Alarms window, and check whether link-related alarms (ALM
22000-22004) are generated for the links between the active and redundancy
USCDBs.

p  2. Check whether the redundancy function is disabled，if it is disabled(the parameter

REDUNDANCYSWITCH has a value of 1, 2 or 5), it means the data consistency
check and(or) data replication are stopped between the active and redundancy
USCDBs, so will occur the symptom.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page37
 Contents
2. DS Principles and O&M
2.1 Data Replication

2.2 Data Verification

2.3 Data Loading

2.4 Data Backup

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page38
Data Loading Principle
DSU/DRU node

Redo log

In-memory Mapping Master

Redo node
Database

Mapping
Data
dictionary
Dump
Loading
files
Insert

DBMS Three ways of restoring data

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page39
 Data Loading Principle (Cont.)
l  The DRU/DSU node can load data from the following four sources in sequence:
p  Local master node

p  Local dump and redolog files

p  Physical database

p  Remote master node

l  This loading mode is used for data loading from the active USCDB to the redundancy
USCDB. This loading mode is disabled by default. The DS adopts the three-level data
backup mechanism.
p  Level-1 backup: The subscriber data is stored in the memories of different VMs. Each cluster is
configured with a master and slave nodes. The data stored in the master node is synchronized to
the slave node in the same cluster in real time.

p  Level-2 backup: The subscriber data stored in the VM memory is backed up to two local hard disks
on the board. The two hard disks work in RAID 1 mode.

p  Level-3 backup: The subscriber data stored in the VM memory is backed up to the disk array. The
disk array is configured with RAID 10 and hot spare disks.
Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page40
 Operations Related to Data Loading
l  Scenarios: Data loading is performed automatically. Data loading sources
changes only in the following scenarios:
p  During the upgrade, the data loading source is automatically changed.

p  After the physical database is cleared up, the data loading source needs to be
changed to load data from the physical database.

p  After the USCDB is restored from abnormal state, the data loading source needs to be
changed to load data from the local dump files and redolog files.

l  If the data loading source needs to be changed in other scenarios, perform the operation
under the guidance of Huawei technical support engineers.

Procedure:
p  Log in to LMT, open the MML command - USCDB window, and run MOD INSP to
modify the value of LDRSOURCE.
p  MOD INSP: INSPT=FE, INSPN=LDRSOURCE, INSPV=X, FETYPEN="USCDB";

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page41
 Operations Related to Data Loading
(Cont.)
l  The values of the LDRSOURCE are as follows:
p  0: Load data from the following sources in sequence: master node,
local dump and redolog files, and the physical database.

p  1: Load data from the master node.

p  2: Load data from the local dump files and redolog files.

p  3: Load data from the physical database.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page42
Solution Related to Data Loading
l  Case: How to locate the fault when a node fails to load data?
Fault Diagnosis:
l  Open the Browse Alarms window, and check the causes for ALM-12007
Loading Failed. The causes for loading failure when a new site is deployed
are as follows:

p  (1) The disk space is insufficient. This situation often occurs in pilot sites
because multiple junk files are generated when NEs are added and
deleted. This cause can be identified based on the Alarm cause in the
alarm.

p  (2) Multiple USCDB versions are tested, and data of the earlier version
in the local dump files is not deleted.

p  (3) Multiple USCDB versions are tested. Data of the earlier version is
deleted but the VMs and NEs are not reset.
Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page43
 Contents
2. DS Principles and O&M
2.1 Data Replication

2.2 Data Verification

2.3 Data Loading

2.4 Data Backup

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page44
 Data Backup Principle
l  The DS provides the following data backup modes:

data backup modes Note

Periodic data backup The backup process starts at 03:38:00 every day to
back up data in the in-memory database to the
local disks.
Manual data backup On HUAWEI Operation & Maintenance System,
run BKP NODE to back up data in the DRU and
DSU nodes.
Automatic backup after After loading data from the physical database, the
loading from the physical DRU or DSU node starts data backup.
database
Automatic backup after After loading data from the active USCDB, the
loading from the remote redundancy USCDB starts data backup.
location

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page45
 Operations Related to Data Backup
(BKP NODE)
l  Scenarios: Perform data backup in the following scenarios:
p  Data migration from Huawei HLR9820 V900R003 or V900R006 to HLRs of other vendors.

p  USCDB capacity expansion.

p  USCDB upgrade.

l  Procedure：Log in to HUAWEI Operation and Maintenance System, open the

MML Command - USCDB window, and run BKP NODE.
p  To back up data in the slave node, set Backup type set to NORMAL. For example, to back
up data in the node 12 in cluster 11.

p  BKP NODE: CID=11, NID=12, BKPTYPE=NORMAL;

p  To back up data in the master node, set Backup type to FORCED. For example, to back
up data in the node 13 in cluster 11.

p  BKP NODE: CID=11, NID=13, BKPTYPE=FORCED;

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page46
 Solution Related to Data Restoration
l  Symptom 1: Data backup for a node fails, and ALM-12007 Loading Failed
is generated with cause value disk space insufficient.

l  Fault Diagnosis : To solve this problem, check the causes for insufficient
disk space and delete useless files.

l  Symptom 2: During the upgrade of a site, the OS of a board is faulty after

the board is upgraded. After the board replacement, no backup files exist on
this board, and data loading fails.

l  Fault Diagnosis : When the standby plane is upgraded, the standby plane
loads data from local dump files. However, the loading fails as the board is
replaced and no backup files exist on the new board. To solve this problem,
copy files from the active plane to the /opt/uscdb/snp_upg directory of the
standby plane and upgrade the standby plane again.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page47
 Questions
1.  If data replication from the master node to the slave node fails,
how the master node ensures data consistency?

2.  When a new site is deployed, where will nodes load initial data?

3.  If the DRU/DSU node restarts, where will the DRU/DSU node
loads data?

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page48
 Contents

1. Overview of the USCDB

2. DS Principles and O&M

3. DBMS (date storage subsystem) Principles and O&M

4. PGW Principles and O&M

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page49
 Contents

3. DBMS (date storage subsystem) Principles and O&M

3.1 Data Synchronization from the In-Memory Database to
the Physical Database

3.2 Data Consistency Check Between the In-Memory and

Physical Databases

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page50
Data Synchronization Principle
1. Static data synchronization
After the static data in the in-

memory database is updated, the
master node in the DRU or DSU
cluster generates the
synchronization log. The
write

synchronization module sends the

In-Memory synchronization log to the DBG
Synchronization log database
module. The DBG module
synchronizes the data to the
1 2
physical database.
read

read

2. Dynamic data synchronization

The DRU or DSU module reads
DBG

data from the in-memory database,
and sends the data to the DBG
write

module. The DBG module

DBMS synchronizes the data to the
physical database.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page51
 Data Synchronization Principle (Cont.)
l  Difference between static data and dynamic data synchronization
p  When the data in the in-memory database is updated, the static data is
synchronized to the physical database in real time. However, the
dynamic data is periodically synchronized to the physical database.

p  The reason is that the performance of the physical database is poorer

than that of the in-memory database. You cannot write and read data in
the physical database quickly and frequently as in the in-memory
database. Therefore, the principle of data synchronization is to
synchronize the most important data to the physical database first. The
data that is restored easily and updated frequently can be
synchronized to the physical database not in a timely manner.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page52
 Data Synchronization
l  Scenarios
p  In normal cases, data is automatically synchronized from the in-memory database
to the physical database. The data synchronization function must be disabled when
physical database upgrade, maintenance, or backup is performed.

l  Procedure
p  Log in to HUAWEI LMT, and run MOD INSP in the MML Command - USCDB
window to enable data synchronization from the in-memory database to the
physical database.

p  MOD INSP: INSPT=FE, INSPN=SYNSWITCH, INSPV=1, FETYPEN="GU-HLR";

n  0 indicates that data synchronization is disabled.

n  1 indicates that data synchronization is enabled.

n  The default value is 1.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page53
 Contents

3. DBMS (date storage subsystem) Principles and O&M

3.1 Data Synchronization from the In-Memory Database to the
Physical Database

3.2 Data Consistency Check Between the In-Memory and

Physical Databases

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page54
Principle of Data Consistency Check Between
the In-Memory and Physical Databases
Slave DSU/DRU node
1.  Based on the primary key, the
DBG queries all data records
in the physical database.
In-memory
database 2.  The DBG compares the
queried data records in the
physical database and the
data stored in the in-memory
database.
1 2 3 3.  If data inconsistency is
detected, data in the physical
DBG database is updated based
on the data in the in-memory
database.
DBMS

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page55
 Principle of Data Consistency Check Between
the In-Memory and Physical Databases (Cont.)
l  Full data verification
p  During each check, the system obtains the data records from the primary key
table of the in-memory database one by one, and then obtains the relevant
records from the corresponding auxiliary key table. Then, the system compares
the records with those stored in the tables of the physical database.

l  Incremental data verification

p  Incremental data verification is enabled in USCDB V100R002C03and later

versions.

p  The speed of data consistency check between the in-memory and physical
databases is slow. Therefore, incremental data verification is used to verify only
the updated data. The last time when subscriber data was updated in the in-
memory database is compared with the time the last data consistency check
between the in-memory and physical databases was performed to determine
whether the subscriber data needs to be updated.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page56
Operations Related to Data Consistency Check
Between the In-Memory and Physical Databases
l  Scenarios
l  In normal cases, data consistency check between the in-memory and physical
databases is automatically performed. Perform data consistency check between
the in-memory and physical databases in the following scenarios:

p  Data synchronization from the in-memory database to the physical database

fails.

p  A redundancy USCDB is deployed.

p  The in-memory database is upgraded by the upgrade tool.

l  Frequently used maintenance commands:
l  1. Run CHK CLUSTER to query the result of the data consistency check between
the physical and memory databases of a cluster.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page57
Operations Related to Data Consistency Check
Between the In-Memory and Physical Databases
(Cont.)
l  2. Run the following command to disable data consistency check between the in-
memory and physical databases on all nodes:

p  MOD INSP: INSPN=IARSTRSTPVFYFLG, INSPV=0;

l  3. Run the following command to enable data consistency check between the in-
memory and physical databases on all nodes:

MOD INSP: INSPN=IARSTRSTPVFYFLG, INSPV=1;

l  4. Run the following command to enable incremental data consistency check between
the in-memory and physical databases on all nodes:

MOD INSP: INSPT=VERIFY, INSPN=IARINCREMENTSWITCH, INSPV=1;

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page58
 Contents

1. Overview of the USCDB

2. DS Principles and O&M

3. DBMS (date storage subsystem) Principles and O&M

4. PGW Principles and O&M

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page59
 Contents
4. PGW Principles and O&M on the PGW
4.1 PGW Right Management (MVNO)

4.3 Subscription Notification

4.4 PGW Logs

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page60
 PGW Right Management Principle
Basic concepts
l  SYSHLRSN(SYSNO): It means System HLR Serial No (System Serial No .).It specifies the system serial number.
Each USCDB is configured with only one system serial number (system serial no ). The system serial number (system
serial no ) identifies a USCDB.

l  VHLRSN(MVNO): This specifies the Virtual HLR Serial No (Mobile Virtual Network Operator ).

l  VHLRRANGE(MVNORANGE): This specifies the number segment that a Virtual HLR(MNVO) manages. Multiple
number segments can be configured for one Virtual HLR(MNVO). The right management of number segments is
performed according to the prefix maximum matching rule. The number segments configured for Virtual HLR(MNVO)
associated with the same system HLR serial number (system serial number )cannot be the duplicate.
SYSHLRSN
(SYSNO)

OPR
VHLRSN1 VHLRSNn
(MVNO1) (MVNOn)

VHLRRANGE1 VHLRRANGE2 VHLRRANGE3 VHLRRANGE4

(MVNORANGE1) (MVNORANGE2) (MVNORANGE3) (MVNORANGE4)

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page61
 PGW Right Management Principle (Cont.)
Basic concepts
l  CCG: This specifies the command group. Commands are grouped. One command group
contains multiple commands, and one command can be included in multiple command
groups.

l  Operator: This specifies the person who can log in to the HLR, configure and manage the
HLR, and manage data stored in the HLR.

p  Operators except the operator admin can only manage data configured for the
system to which the operators belong. The system is identified by the SYSNO.

p  Right management is aimed at controlling the operations of operators who log in to

the PGW system.

l  Operator admin: In the PGW system, one operator admin is configured by default. The
operator admin has full administrative rights.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page62
PGW Right Management Principle (Cont.)
Overview
l  The PGW supports the following right management:
p  Command right management
p  Number segment right management
l  PGW right management is implemented based on operator right
management. The operator rights include:
p  Command right
p  Number segment right
l  The number segment right management is implemented by configuring
VHLRSNs(MVNOs) for an operator and number segments for the
VHLRSNs(MVNOs).

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page63
PGW Right Management

l  Command right management

l  Number segment right management

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page64
 Command Right Management Principle
The principle of command right management is as follows:
l  The commands that an operator has the right to run is configured by setting
the operator level.
l  The operator level is associated with the command group.
l  The command group can be configured by running ADD CCG.
l  An operator has the right to run all commands in the command group
associated with the operator level.

Operator 1 Level 1
Command 1
Level 2 Command 2
Command group 1
Operator 2
Level 3 Command group 2 Command X
Operator X
Command group X
Level X

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page65
Command Right Management Principle (Cont.)
The predefined command groups are as follows:

Command Group ID Function Description

G_0 This command group is used to query

subscriber data.
G_1 This command group is used to modify
subscriber data.
G_2 This command group is used to query
global data or template data.
G_3 This command group is used to modify
global data or template data.
G_4 This command group is used to query
authentication data.
G_5 This command group is used to modify
authentication data.
G_6 This command group is used to query
service data.
G_7 This command group is used to modify
service data.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page66
Command Right Management Principle (Cont.)

Operator level
l  Operators are classified into five levels as follows:
p  ADMINISTRATOR
p  OPERATOR
p  USER
p  GUEST
p  CUSTOM

l  The command groups associated with the first four operator levels
are predefined. You can assign customized command groups to an
operator of the Custom level if required.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page67
 Command Right Management Principle (Cont.)
The mapping between operator levels and command groups is as follows:

Operator Level Command Group

GUEST G_0, G_2, G_4, and G_6
USER G_0, G_1, G_2, G_4, G_6, and G_7
OPERATOR G_0, G_1, G_2, G_3, G_4, G_5, G_6, and G_7
ADMINISTRATOR G_0, G_1, G_2, G_3, G_4, G_5, G_6, G_7, security management
commands, operator management commands, command group
management commands, and operator-oriented MVNO management
commands
admin G_0, G_1, G_2, G_3, G_4, G_5, G_6, G_7, security management
commands, operator management commands, command group
management commands, SYSHLRSN(SYSNO) management
commands, Virtual HLR(MVNO) management commands, Virtual
HLR(MVNO) number segment management commands, and operator-
oriented MVNO management commands
CUSTOM Command groups between G_0 and G_127 with predefined commands

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page68
Operations Related to Command Right Management
Prerequisites
l  The HLR serial number must be added before operators are added.
Procedure
l  1. Log in to the PGW Web LMT, and run ADD CCG to add command groups.

ADD CCG: SYSHLRSN=1, CG=G_8, CMD=LST_LOG;

ADD CCG: SYSNO=1, CG=G_8, CMD=LST_LOG; (display MVNO feature)

l  2. Set the name of command groups.

SET CCGN: SYSHLRSN=1, CG=G_8, CGNAME="NormalUser";

SET CCGN: SYSNO=1, CG=G_8, CGNAME="NormalUser"; (display MVNO

feature)

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page69
 Operations Related to Command Right
Management (Cont.)
l  3. Run ADD OPR to add operators.

For example, to add an operator on the system whose serial number is 1,where:
Operator name is pgwuser
Operator description is user in pgw system
Password is Abc123
Operator level is User
Operator status is ACTIVE

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page70
Example of Configuring Command Right
Management
l  In normal cases, one or two operators of the Administrator level and
multiple operators of the User level are configured for the PGW.
p  An operator of the Administrator level can manage global data, such as
template customization.

p  An operator of the User level can add or delete subscribers, provide

services, or query global data.
n  To add an operator of the Administrator level, run the following command:

ADD OPR: SYSNO=1, OPNAME="opradmin1", PWD="*****", CFM="*****",

UG=ADMINISTRATOR, STATUS=ACTIVE;

n  To add an operator of the User level, run the following command:

n  ADD OPR: SYSNO=1, OPNAME="opruser1", PWD="*****", CFM="*****",

UG=USER, STATUS=ACTIVE;
n  You can query rights of an operator only by running LST OPR.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page71
PGW Right Management

l  Command right management

l  Number segment right management

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page72
Number Segment Right Management
Maximum matching rule

l  Number segment maximum matching is used to obtain the number segment that
best matches the number of a specified subscriber. The procedure is as follows:

p  1. The USCDB obtains all number segments in the system.

p  2. The USCDB matches the number of the subscriber with the number
segments and obtain the number segment that best matches the number and
the mapping VHLRSN(MVNO).

p  3. The USCDB queries all manageable Virtual HLRs(MVNOs) for the current
operator.

p  4. The USCDB determines whether the VHLRSN(MVNO) obtained in step 2 is

included in the manageable VHLRSNs(MVNOs). If the VHLRSN(MVNO) is
included, the operator has the right to manage the number segment. If not, the
operator does not have the right to manage the number segment.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page73
Operations Related to Number Segment
Right Management Virtual HLR Management

Virtual HLR commands

l  Virtual HLR right

management for operators
is implemented by
configuring manageable
Virtual HLRs for the
operators.

l  Only the operator admin

has the right to run ADD
SYSHLRSN.

l  The other operators only

have the right to manage
subscriber data managed
by the Virtual HLR
configured for the
operators.
Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page74
Operations Related to Number Segment
Right Management (Cont.)
Operator-Oriented Virtual
HLR Management

l  An operator of the
Administrator level has the
right to add manageable
virtual HLRs for other
operators.

l  Multiple manageable virtual

HLRs can be added for an
operator.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page75
Operations Related to Number Segment
Right Management (Cont.) Virtual HLR Number
Segment
Management
l  Number segment right
management for
operators is
implemented by
configuring number
segments for virtual
HLRs and
manageable virtual
HLRs for the
operators.
l  The number
segments that one
virtual HLR can
manage are planned
by carriers.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page76
Operations Related to Number Segment
Right Management (Cont.)
Mobile Virtual Network
Operator Management

MVNO management commands

l  MVNO right management for

operators is implemented by
configuring manageable
MVNOs for the operators.

l  Only the operator admin has

the right to run ADD SYSNO.

l  The other operators only

have the right to manage
subscriber data managed by
the MVNOs configured for the
operators.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page77
Operations Related to Number Segment
Right Management (Cont.)
Operator-Oriented Mobile Virtual
Network Operator Management

l  An operator of the Administrator

level has the right to add
manageable MVNOs for other
operators.

l  Multiple manageable MVNOs can

be added for an operator.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page78
Operations Related to Number Segment
Right Management (Cont.)
Mobile Virtual Network Operator
Number Segment Management

l  Number segment right management

for operators is implemented by
configuring number segments for
MVNOs and manageable MVNOs for
the operators.

l  The number segments that an MVNO

can manage are planned by carriers.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page79
 Example of Configuring Number Segment
Right Management
Background
l  Carriers K, T, M manage subscriber data by using the same set of SingleSDB (HLR)
devices as follows:
p  Carrier K manages the IMSI number segments with number segment with number
prefixes 20408, 20412107 and 20412807, excluding the number segments with
number prefixes 2040807, 2040809, and 20408080.
p  Carrier T manages the IMSI number segment with number prefix 20412, excluding
the number segments with number prefixes 20412107 and 20412807.
p  Carrier M manages the IMSI number segment with number prefix 20408080.

l  Based on the preceding information, the following three operators of the Administrator
level are required:
p  Operator K can manage MVNO1.
p  Operator T can manage MVNO2.
p  Operator M can manage MVNO3.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page80
 Example of Configuring Number Segment
Right Management (Cont.)
Procedure for configuring MVNO right management
l  Note:

p  The added SYSNO must be consistent with that configured

by running SET SYSNO on HUAWEI Operation &
Maintenance System.

p  Run ADD MVNORANGE to add number segments that the

MVNO manages.

p  In the example, run ADD MVNORANGE to add the IMSI

number segments with number prefixes 20408, 20412107,
and 20412807 for MVNO1, the IMSI number segment with
number prefix 20412 for MVNO2, and the IMSI number
segment with number prefix 20408080 for MVNO3.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page81
 scripts of Configuring Number Segment
Right Management (Cont.)
Configuring MVNOs
p  Configure an SYSNO. (The SYSNO must be consistent with that configured by running SET SYSNO on LMT.)

ADD SYSNO: SYSNO=1, STATUS=MASTER;

p  Configure MVNOs. (MVNOs 1, 2, and 3 map carriers K, T, and M. MVNO 4 is used to exclude the unnecessary
number segment for carrier K.)

ADD MVNO: SYSNO=1, MVNO=1, MVNONAME="K";

ADD MVNO: SYSNO=1, MVNO=2, MVNONAME="T";

ADD MVNO: SYSNO=1, MVNO=3, MVNONAME="M";

ADD MVNO: SYSNO=1, MVNO=4, MVNONAME="other";

Configuring number segments

l  Run ADD MVNORANGE to add number segments for the MVNOs.

p  ADD MVNORANGE: SYSNO=1, MVNO=1, RANGETYPE=IMSI, PREFIX="20408";

p  ADD MVNORANGE: SYSNO=1, MVNO=1, RANGETYPE=IMSI, PREFIX="20412107";

p  ADD MVNORANGE: SYSNO=1, MVNO=1, RANGETYPE=IMSI, PREFIX="20412807";

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page82
 scripts of Configuring Number Segment
Right Management (Cont.)
p  ADD MVNORANGE: SYSNO=1, MVNO=2, RANGETYPE=IMSI, PREFIX="20412";

p  ADD MVNORANGE: SYSNO=1, MVNO=3, RANGETYPE=IMSI, PREFIX="20408080";

p  ADD MVNORANGE: SYSNO=1, MVNO=4, RANGETYPE=IMSI, PREFIX="2040807";

p  ADD MVNORANGE: SYSNO=1, MVNO=4, RANGETYPE=IMSI, PREFIX="2040809";

Configuring Operators

l  Configure operators.

p  ADD OPR: SYSNO=1, OPNAME="OPRK", PWD="*****", CFM="*****", UG=ADMINISTRATOR, STATUS=ACTIVE;

p  ADD OPR: SYSNO=1, OPNAME="OPRT", PWD="*****", CFM="*****", UG=ADMINISTRATOR, STATUS=ACTIVE;

p  ADD OPR: SYSNO=1, OPNAME="OPRM", PWD="*****", CFM="*****", UG=ADMINISTRATOR, STATUS=ACTIVE;

l  Configure manageable MVNOs for the operators.

p  ADD OPRMVNO: SYSNO=1, OPNAME="OPRK", MVNO=1;

p  ADD OPRMVNO: SYSNO=1, OPNAME="OPRT", MVNO=2;

p  ADD OPRMVNO: SYSNO=1, OPNAME="OPRM", MVNO=3;

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page83
 Questions
l  Can one IMSI number segment be allocated to multiple operators?

l  How to configure data to ensure that an operator cannot manage the

number segment managed by other operators, for example, operator K
cannot manage the IMSI number segment managed by operator T?

l  How to configure operators for MVNO Telfort and configure number

segments for the added operators?

l  How to configure data to ensure that a specific operator can only run
commands to query subscriber data or commands in a specified
command set?

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page84
 Contents

4 PGW Principles and O&M on the PGW

4.1 PGW Right Management(MVNO)

4.2 Subscription Notification

4.3 PGW Logs

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page85
 Subscription Notification
Implementation Principle

l  If a service is provided or canceled on the provisioning system or

PGW Web LMT, the PGW notifies the network-side network elements
(for example, the MSC server or SGSN) of the change through the
FE.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page86
Subscription Notification (Cont.)
Workflow
1. The provisioning system sends a request for modifying the subscriber data over the
MML or SOAP interface to the PGW.

2. The PGW sends a message to the USCDB to modify the subscriber data.

3. After the subscriber data is modified, the USCDB return a success response to the
PGW.

4. The PGW determines whether to send a notification based on the subscription data
to the FE.

5. The FE analyzes the notification content (in XML format), and sends an MAP
notification to the MSC server.

6. The MSC returns a response to the FE.

7. The FE return a response to the PGW.

8. The PGW returns a success response to the provisioning system.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page87
Subscription Notification (Cont.)
Mechanism for common subscription
notification
FE1 FEn
Notify (C, l  The FE subscribes to notifications
OldValue, based on data models.
NewValue
l  The PGW sends a notification to the
) IP network Modify FE when the subscription data is
(C) modified.
l  The notification contains the data
before and after the modification.
BE1 BE2 PGW
You must specify Location ID if the
l 

UPCC supports notification sending

to multiple FEs.
The subscription notification function ensures that the FEs obtain the updated data
in time and determines whether to update network data based on the information
contained in notifications.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page88
 Subscription Notification (Cont.)
Mechanism for third-party subscription notification

FE1 FEn
Notify (C, Third-Party server
OldValue,
NewValue)
IP network

Modify (C)
BE1 BE2 PGW

In the mechanism for third-party subscription notification, the FE transfers the notification
sent by the FE to a third-party server for data collection and analysis after the dynamic
and static data of subscribers is modified. The third-party subscription notification
configured by ADD NOTIFYCFG does not affect the common subscription notification.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page89
Subscription Notification (Cont.)
Classification of subscription notification

l  Based on the function, the subscription notification is classified into:

p  Subscription update notification

p  Notification update notification

l  Based on the content or workflow, the subscription notification is classified into:

p  Global data notification

p  Subscriber data notification

l  Based on the interface, the subscription notification is classified into:

p  Notification over the SOAP interface

p  Notification over the MCI interface

p  Notification over the FILE interface when the authentication HLR (AUC) feature is
used or the PGW forwards processing results of the requests sent from the
provisioning system to a third-party device

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page90
Subscription Notification (Cont.)
Configuring the subscription notification function
l  Run ADD NOTIFYCFG to configure the notification subscription data on the PGW.

p  To add a notification subscription data record, where:

l  FE type is GU-HLR
l  Remote ID is 1
l  FE ID is 3
l  Interface type is MCI
l  Synchro switch is SYNC

p  run the following command:

ADD NOTIFYCFG: FETYPE="GU-HLR", REMOTEID=1, FEID=3, INFTYPE=MCI,

SYNCHROSWITCH=SYNC;

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page91
 Solution Related to Subscription
Notification
l  Symptom: The subscription notification sending failure (In the command output, the
result shows that data is successfully updated in the database, but data fails to be
updated at the network side.)

l  Fault Diagnosis：

1. The peer FE is faulty.

2. The DSG links between the BE and the FE is faulty.

3. The parameters of the ADD NOTIFYCFG command are incorrectly configured, for
example, Remote ID is incorrectly configured. (The value of Remote ID is set
before product delivery.)

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page92
 Contents

4. PGW Principles and O&M

4.1 PGW Right Management(MVNO)

4.2 Subscription Notification

4.3 PGW Logs

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page93
PGW Logs

Overview
l  The PGW logs are classified into:
p  PGW Operation logs

p  PGW processes logs

p  Operation logs of the PGW Web LMT

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page94
 PGW Operation Logs
Overview
l  The operation logs record all non-query commands executed on the PGW Web LMT and the
provisioning system. The logs are stored in the physical database of both the active and redundancy
USCDBs.

l  By default, the operation logs record operations in latest 31 days. You can configure the storage
duration by setting PGWOPLOGCYC. The value of PGWOPLOGCYC ranges from 1 to 31.

l  To query the value of PGWOPLOGCYC, run the following command:

LST INSP: INSPT=PGW, INSPN=PGWOPLOGCYC;

PGW

DBMS

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page95
PGW Operation Logs(cont.)
Provisioning
Signaling access BSG/DPU BSG/DPU … system
FE
Service logic CCU CCU …

Data routing DRU DRU DRU

… DRU DRU DRU PGW 1
. PGW N
cluster1
cluster N

BE Data service Crash!

DSU DSU DSU DSU
… DSU DSU
DSU cluster 1 DSU cluster 2 DSU cluster N

Data storage Physical database

The operation logs can be used for:

l  DRU or DSU data restoration from the physical database and operation logs in the
service layer redundancy solution.
l  Reconciliation of the provisioning system.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page96
 PGW Operation Logs(cont.)
Redundancy Synchronization file
Redundancy Synchronization file generation
l  After a command is successfully executed, the command is recorded in the redundancy file.

l  The redundancy synchronization file is named in

HLRXXX_YYYYMMDD_HH_MM_SS_NNNNNNNNNN_SYN.DAT

l  The redundancy synchronization file is stored in the /opt/dpu/redundance/realtime/ directory of

the PGW VM where the active DPU resides.

Redundancy file transfer

l  The active DPU on the PGW VM of the redundancy USCDB obtains the redundancy file stored in
the realtime and takeover directories of the active USCDB, store the redundancy file in the syn
directory, and delete the redundancy file on the active USCDB.

p  Note: If the service layer redundancy solution is not adopted, log files are not generated in the
preceding directories.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page97
 PGW Operation Logs(cont.)
/opt/pgw

| |-- ./redundance/pgw

| | |-- ./redundance/pgw/realtime (the redundancy file generated on the active USCDB)

| | |-- ./redundance/pgw/syn (the redundancy file generated on the redundancy USCDB)

| | |-- ./redundance/pgw/takeover (the synchronization file )

Redundancy file execution

l  The USCDB scans the redundancy files stored in the syn directory. If a redundancy file that meets
requirements are found, the USCDB performs a batch operation to execute the redundancy file. The
interworking between the active and redundancy USCDB may be temporarily abnormal or the
redundancy file is not transferred in real time in some cases. Responding to it, the redundancy file is not
executed in real time. The USCDB only executes the synchronization file generated 5 minutes ago.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page98
PGW Processes Logs
Overview

l  The logs of PGW processes are stored in the /opt/ne/MEID/proc/

workspace0/prog/ directory of the PGW VM.

p  Initialization log : It is named in the format of UscInit+Module number

of the process.log, for example, UscInit3.log.

p  Local debugging log : It is named in the format of UscDebug+Module

number.log, for example, UscDebug3.log.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page99
 Solution Related to PGW Processes Logs
Symptom: PGW Web LMT login failures.

Fault Diagnosis:

l  If the login page of the PGW Web LMT cannot be displayed, perform the
following operations to check whether the IP address configured for the
PGW is correct.

p  In the MML Command - USCDB window, run LST IPADDR to query the
IP address configured for the PGW.

p  Check whether the IP address is pingable. If it is not pingable, check the

network connection. If it can be pingable, contact Huawei technical
support engineers.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page100
Operation Logs of the PGW Web UI
l  Log in to the PGW Web LMT, and click the Operation Record tab. The MML
commands that have been executed recently are displayed.

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page101
 Questions

l  How operator rights are managed?

l  What function does the subscription notification mainly
provide?

l  Which directories of the PGW VM for data synchronization

are used in the service layer redundancy solution?

Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page102
 Summary

This course describes the followings:

l  Working principles of the DS, DBMS, and PGW modules.

l  Mechanisms of data replication, verification, loading, and

backup and O&M on the DS.
l  Mechanisms of data synchronization from the in-memory
database to the physical database and data consistency check
between the in-memory and physical databases and O&M on the
DBMS.

l  MVNO right management, subscription notification, PGW

logs, and O&M on the PGW.
Copyright © 2017 Huawei Technologies Co., Ltd. All rights reserved. Page103
Thank you!
www.huawei.com