Guide to Clearing the DBMS_ERROR Alarm Reported After an Upgrade Due to the Residual SNCP Pointe
Constraint: This recovery guide is applicable to V200R011C03SPC200, V200R012C00,
and V200R012C01 only
1. Identifying the Problem
Run the :dbms-query:"23310000.dbf",mdb command. The query result is shown in the
following attachment: “Result.txt”
As shown in the attachment in 1.1 "Querying the Database", the 3112103 field is 00 but
the 23312301 field is not 00000000, so the SNCP pointer exists.
Records 1, 2, 5, and 6 need to be recovered, as shown as follows:
2. Recovering the Database
2.1 Identifying the Recovery Mode
2.1.1 V200R011C03SPC200, V200R012C00, or V200R012C01
alm-get-curdata-ext:0,DBMS_ERROR,0; command to check whether the DBMS_ERROR
alarm is reported.
If the DBMS_ERROR alarm is not reported, perform the steps in 2.2 "Recovery Method
When No DBMS_ERROR Alarm Is Reported"; if the DBMS_ERROR alarm is reported,
perform the steps in "Recovery Method When the DBMS_ERROR Alarm Is Reported".
2.1.2Versions Earlier than V200R011C03SPC200
check whether the DBMS_ERROR alarm will be reported after the upgrade. If the
DBMS_ERROR alarm is not reported, perform the steps in "Recovery Method When No
DBMS_ERROR Alarm Is Reported" after normally upgrading the NE; if the DBMS_ERROR
alarm is reported, the activation of the processing board will time out, so perform the
steps in "Recovery Method When the DBMS_ERROR Alarm Is Reported".
is as follows:
In the query result, the record of the residual SNCP pointer exists. If the value of the 23312301
field is displayed twice, the DBMS_ERROR alarm will be reported after the upgrade; if not,
the DBMS_ERROR alarms will not be reported.
The query result in 1.1 "Querying the Database" is used as an example. Confirm that records 5
and 6 are residual data by referring to 1.2 "Checking Whether a Residual SNCP Pointer Exists".
The value of the 23312301 field is c7240000. In the query result, c7240000 is displayed twice.
The DBMS_ERROR alarm will be reported.
2.2.1 Running Commands to Recovery the Database
:cfg-gsp-test:"modifyPO 23310000 poid 23312301 00000000"
the residual SNCP pointer records. Values of other fields are fixed (pay attention to the space
between data).
and 9323 respectively.
Run the following two commands:
:cfg-gsp-test:"modifyPO 23310000 9321 23312301 00000000"
:cfg-gsp-test:"modifyPO 23310000 9323 23312301 00000000"
Recover all residual records in sequence.
2.2.2 Checking Recovery Results
Run the :dbms-query:"23310000.dbf",mdb command to confirm that all residual SNCP pointer
records are recovered.
2.2.3 Saving the Database
Run the :dbms-copy-all:drdb,fdb command.
The recovery is complete.
2.3.1 Temporarily Allowing External Operations
Run the :sm-set-nebusy:0,0,0,0,none command.
When the DBMS_ERROR alarm is reported, no external operation is allowed on
the NE. After the command is executed, you can perform operations on the NE
in five minutes. If the residual records are not recovered after five minutes, you need to run
the command again. Do not configure the NE during this period.
2.3.2 Running Commands to Recovery the Database
:cfg-gsp-test:"modifyPO 23310000 poid 23312301 00000000"
Poid indicates the decimal value converted from the hexadecimal value of
the 14112301 field in the residual SNCP pointer records. Values of other
fields are fixed (pay attention to the space between data).
For records 1 and 2, the decimal values for 2469 and 246b in the 14112301
field are 9321 and 9323 respectively.
2.3.Run the following two commands:
:cfg-gsp-test:"modifyPO 23310000 9321 23312301 00000000"
:cfg-gsp-test:"modifyPO 23310000 9323 23312301 00000000"
Recover all residual records in sequence.
Backup
After completing all recovery operations, wait for about 45 minutes. Run
the :alm-get-curdata-ext:0,DBMS_ERROR,0; command to confirm that the DBMS_ERROR alarm is cleared.
Run the :hbu-get-backup-info command and confirm that the returned value is 3.
The recovery is complete.
If the returned value is not 3, contact Huawei's engineers.
2.4 Upgrade and Recovery Measures for the NE on Which the DBMS_ERROR
Alarm Is Reported
After an Upgrade
The upgrade steps are adjusted. The system control board and processing board
are loaded and activated respectively. Upgrade the system control board. After
the fault is rectified, upgrade the processing board.
Simulation Package Loading
When confirming the simulation loading and activation sequence, for the group
where the system control board is not located, select Pause Before Current
Group, as shown as follows:
After the system control board upgrade is complete, the NE stops the activation
of more boards.
2.4.2 Temporarily Allowing External Operations
Run the : sm-set-nebusy: 0,0,0,0,none command.
External operations are not allowed on the NE during backup. After the command
is executed, you can perform operations on the NE in about five minutes. If the
residual records are not recovered after five minutes, you need to run the
command again. Do not configure the NE during this period.
2.4.3 Running the Command to Recovery the Database
:cfg-gsp-test:"modifyPO 23310000 poid 23312301 00000000"
Poid indicates the decimal value converted from the hexadecimal value of the
14112301 field in the residual SNCP pointer records. Values of other fields
are fixed (pay attention to the space between data).
For records 1 and 2, the decimal values for 2469 and 246b in the 14112301
field are 9321 and 9323 respectively.
Run the following two commands:
:cfg-gsp-test:"modifyPO 23310000 9321 23312301 00000000"
:cfg-gsp-test:"modifyPO 23310000 9323 23312301 00000000"
Recover all residual records in sequence.
2.4.4 Waiting for Alarm Clearance and Completing Active/Standby
Backup
After about 45 minutes, run the :alm-get-curdata-ext:0,DBMS_ERROR,0; command to confirm that the DBMS_ERROR alarm is cleared.
Run the :hbu-get-backup-info command and confirm that the returned value is 3.
If the returned value is not 3, contact Huawei's engineers.
2.4.5 Continuing Tasks and Completing Processing Board Activation
The recovery is complete.
More blog:
No comments:
Post a Comment