I. Description
The M2000 client fails to connect to the alarm server. It is displayed on the RWS that the connection to the alarm server is interrupted or abnormal.
II. Analysis
When the client fails to connect the alarm server, check whether the alarm server process is normal and whether the number of alarm records exceeds the limit in the alarm database.
Note:
When the number of history alarms exceeds the preset value, the M2000 alarm server deletes redundant history alarms. That is, the M2000 alarm server deletes at most 5000 alarms per type per level once every half an hour.
When the system generates large amount of alarms owing to upgrade or other reasons, the alarms reported in half an hour may be more than alarms the system deletes automatically. If this situation lasts, the history alarms will be over redundant, thus affecting the normal operation of the alarm server process.
III. Solution
To handle the problem, proceed as follows:
1) Verify whether the version of the M2000 common service is B03D501 or earlier. YÖ 2)
NÖ Contact Huawei technical support engineers.
2) Check whether history alarms are over redundant in the M2000 alarm database. Proceed as follows:
Log in to the M2000 server as m2000.
Execute the following commands to query the history alarm table. $ isql -Usa -Pserver1234
1> use alarmdb 2> go
1> select count(*) from m2000.history /*Check the number of history alarms*/ 2> go
512345 (1 row affected)
If the number of records is over 500,000, it is recommended to clear all the history alarms.
If the number of records is small, directly go to step 5.
Confirm whether all the history alarms can be deleted because the deletion is irretrievable.
Confirm whether to delete all the history alarms. YÖ 3)
NÖ 4)
3) Clear all the records of history alarms in the M2000 alarm database. Proceed as follows:
Log in to the M2000 server. Stop the M2000 server process. $su
#/etc/rc2.d/S98m2000svr stop
Execute the following commands to clear the history alarm table. #su – m2000
$isql -Usa -Pserver1234
1> use alarmdb 2> go
1> truncate table m2000.history 2> go
1>exit
Restart the M2000 server process.
$su
#/etc/rc2.d/S98m2000svr start
Check whether the problem is solved five minutes later. YÖ End.
NÖ 5)
4) If the deleting of all the history alarms is not permitted, you can delete the history alarms generated before a specific time, which needs a longer time. To delete
history alarms generated before a specific time, please execute the following commands:
Log in to the M2000 server. Stop the M2000 server process. $su
#/etc/rc2.d/S98m2000svr stop
Execute the following command to delete history alarms generated before a specific time.
#su – m2000
$isql -Usa -Pserver1234
1> use alarmdb 2> go
1> set rowcount 5000 2> go
/*Execute the following commands repeatedly until the system does not display “(5000 rows affected)”. */
1> delete from m2000.history where occurtime<"20040301 00:00:00"
2> go
/*The above commands mean all the history alarms generated before 00:00:00 on March 1, 2004 will be deleted. You may change the time as needed. */
Restart the M2000 server process. $su
#/etc/rc2.d/S98m2000svr start
Check whether the problem is solved five minutes later. YÖ End.
NÖ 5)
5) Check whether event alarms are over redundant in the M2000 alarm database. Proceed as follows:
Log in to the M2000 server as m2000.
Execute the following commands to query the event alarm table. $ isql -Usa -Pserver1234
1> use alarmdb 2> go
1> select count(*) from m2000.event /*Check the number of event alarms*/ 2> go
512345 (1 row affected)
If the number of records is more than 500,000, it is recommended to clear all the event alarms.
If the number is small, directly go to step 8.
Confirm whether all the event alarms can be deleted before deleting them. (This operation is irretrievable. )
YÖ 6) NÖ 7)
6) Clear all the event alarms in the M2000 alarm database.
Change the table name m2000.history to m2000.event and proceed as the step 3 describes.
Check whether the problem is solved five minutes later. YÖ End.
NÖ 8)
7) If the deleting of all the event alarms is not permitted, you can delete the event alarms generated before a specific time. However, it will take comparatively longer time.
Change the table name m2000.history to m2000.event and proceed as the step 4 describes.
8) Proceed with the following operations Log in to the M2000 server as m2000.
Execute the following commands to query the number of alarm records. $ isql -Usa -Pserver1234
1> use alarmdb 2> go
1> select count(*) from m2000.history /*Check the number of history alarms*/ 2> go
1> select count(*) from m2000.fault /*Check the number of fault alarms*/ 2> go
1>select count(*) from m2000.event /*Check the number of event alarms*/ 2>go
1>exit
Contact Huawei technical support engineers and feed back the returned result of the above commands and the log files under the directory /export/home/m2000/log.
4.2.9 Failed to Report Alarm Data
I. Analysis
The possible causes are:
z The server does not report alarms. z The alarm server is abnormal. z The database is abnormal.
II. Solution
To handle this problem, check whether:
z The NE version is consistent with the adapter version. z There is certain alarm on LMT of host.
z Alarms are shielded.
z Alarms are repeated alarms. z Alarms are newly added.
z The server works normally if all NEs cannot report alarms.
z The NE adapter works normally if all alarms of a certain NE cannot be reported.