Problem description

M operator OSN3500 configures IDQ1 single board, when IDQ1 single board is added to the T2000 of the network management, GSCC immediately soft resets (on-site observation and querying the master errlog ), and the network element is grayed out on the network management; the network element automatically logs in after 1~2min, and checks the status at this time: the IDQ1 single board has been added. After that, when doing IDQVctrunk binding operation, GSCC soft reset at the same time of issuing commands, and command line operation will also lead to GSCC automatic soft reset. The version of the main control board is 5.21.13.47.

Warning message

Main control board auto soft reset

Processing

1. Since the main control board will automatically soft reset only when binding Vctrunk to IDQ, other operations such as configuring SDH cross-connections are not a problem. Initially, it is suspected that there is a problem with the database of the network element Vctrunk.

2. Analyzing the reset record of the main control board, we found that the automatic reset of the main control board is indeed caused by the operation of the ATM database. Further analysis of the network element database file found that the database version of the problem network element and the database version of the normal network element are inconsistent, which led to database read and write failures, the main control reset.

3, the site feedback of the previous version of this network element to do the upgrade, from 5.21.12.43 upgrade to 5.21.13.47,

Since the database versions of these two versions are different, it is certain that it is due to the upgrade that the database version inconsistency problem occurs.

4. By analyzing the returned dbmslist.inf file, we found that the version of atm_trk.dbf is still V1. In other words, atm_trk.dbf was not upgraded successfully during the master upgrade process. Through local testing, it was found that when the host version of atm_trk.dbf for R2 is an empty library, the upgrade will cause the database version conversion failure, which will lead to the problem. Detailed analysis of the process can refer to the attachment.

5, the network element to wipe the library hard reset and redo the configuration, or the normal network element atm_trk.dbf database file copied to the faulty network element and then soft reset the master can be restored in an emergency.

Root Cause

When the host version is R2, if atm_trk.dbf is an empty database, the upgrade will cause the database version conversion to fail, which in turn will cause the database read/write to fail, and the master board will automatically reset.

Suggestions and Summary

This issue only occurs when atm_trk.dbf is an empty library. Before upgrading from V1R2 version to V1R3 and subsequent versions, you need to query the ATM 's vctrunk database atm_trk.dbf via Navigator to determine whether the problem will occur if atm_trk.dbf is empty (use the command: :dbms-query: "atm_trk.dbf",mdb ). If it is empty (returns NSERR_DBMS_DBF_IS_EMPTY ), you need to initialize the database using the atm-init command (at this point, you need to reconfigure the ATM service).