No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

New Cross-Connection Cannot Be Set Up due to Database Exceptions Occurred After the OptiX OSN 3500 Upgrade

Publication Date:  2012-07-25 Views:  56 Downloads:  0
Issue Description
A user configures the VC-12 circuit through the T2100, but the configuration fails. The system prompts that errors occur in the system function call of the OptiX OSN 3500 in the 6055-Yuquan Road Trail.
The engineers set the cross-connection of the OptiX OSN 3500 to the idle timeslot through the timeslot selection in the T2100. After multiple attempts, the system still prompts error in the system function call. The engineers also try in the T2000, but the attempts still fail. The system prompts that “Conflict at the sink of the cross-connection service and errors occur in the system function call”. The engineers query the version of the OptiX OSN 3500 and find that the version is 13.47 and the version is upgraded from an earlier version. 
 
Alarm Information
Error occurs to system function call and the sink of the cross-connection service conflicts. 
Handling Process
The engineers analyze the NE data and conclude that the sink cross-connection conflict in configuration is caused by the overlapped subtree mounting relationships of boards in the database of the SCC. The engineers decide to repair the database of the SCC by downloading all the configuration data through the T2000V2R3C01 in non-interrupt mode.
The engineers operate the NE in the field. Before the download, the engineers confirm that the versions of the active and standby SCCs are the same, the active and standby SCCs work normally, the batch backup is normal, and no abnormal alarms are reported. After the engineers download the configurations, the T2000 prompts the NE initialization, but the progress bar displayed in the dialog box does not proceed. The engineers check the board status and find that all indicators on the active SCC are off, but the indicators are lit after a time. Based on pervious experiences, the engineers conclude that the active SCC is reset. After about 10 minutes, the system prompts the download failure: “The server is busy and the NE initialization fails. The MSP ring exists and the operation on MSP ring fails”.
The engineers check the operation status of the OptiX OSN 3500 immediately and find that the NE is in the installation state. The engineers also check the alarm and MSP ring status and find that slots 8 and 11 of the NE are mapped to a MSP ring and both the eastern and western directions of the MS are switched. The engineers disable the protocol of the entire network, but the NE is still in the switching state. Because the optical powers at the two ends are normal, the engineers exclude the possibility that the symptom is caused by a normal switching. The engineers check the K byte immediately and find that the NE sends the switching request to both the ends. As a result, the switching occurs.
The engineers conclude from the symptoms that, because the database of the active SCC is abnormal, the active SCC is reset immediately when the configuration is downloaded from the NE. After the active SCC starts to work, the active and standby SCCs are synchronized automatically. As a result, the download fails. Before the download, part configuration data in the MS of the SCC is cleared. Hence, the MS protocol fails.
The engineers directly remove the active SCC and the download restores normal. 
 
Root Cause
In the upgrade from the R2 to the R3, the real-time database storage function is suspended when the nvram is full. A reset is performed on the board because of the change of working/protection status. As a result, part files are not backed up to the flash in time. After the board is reset, the database recovery fails. The MO created later occupies the PrevGroupId recorded previously. As a result, the boards in slot 5 and slot 14 are mounted abnormally. Consequently, when the board in slot 5 is set to the sink, the subtree of the board in slot 14 is found and the returned value displays that the cross-connection exists.
The subsequent downloading of the configuration data fails. The reason is that the MO That is incorrectly mounted to the tree is abnormal in the initialization. As a result, the NE is reset and the downloading fails. However, the optical port protection type of the MSP in the line board has been set to null. Because the NE software enables the MSP protocol again after the NE is reset, exceptions occur to the MSP switching and the exception cannot be cleared by enabling/disabling the MSP protocol. 
 
Suggestions
In the operations on the current network, ensuring the running of services should be of the greatest concern. For the plans developed by R&D, the most secure and feasible plan should be used.
To avoid the problem caused by database incompleteness as in this case, a database completeness check mechanism is introduced to the R7. When the drdb is found to be incomplete, the database is restored from the fdb. When both the fdb and drdb databases are checked as incomplete, the system actively clears the databases, enters the installation state, and reports the installation state alarm.
The engineers can also delete databases drdb, fdb0, fdb1, and nvfldb by using certain commands, and delete the files ocp/mover, /ocp/dbfver, /ocp/mofile, and /ocp/moptr in the mfs, ofs1, and ofs2 by running the command sftm-delete-file. Then, the engineers should reset the SCC and re-download the configuration. 
 

END