For a set of OSN 3500 with an extended subrack, the service of its extended subrack is interrupted suddenly, with alarms of LP_SIZE_ERR and TU_AIS, which continue for about two minutes each time. The NE host software version is 220.127.116.11.
772817 54 LP_SIZE_ERR CR end 2007-5-14 23:4:12 2007-5-14 23:6:27 0x01 0x00 0x36 0xff 0x01
772817 54 LP_SIZE_ERR CR end 2007-5-14 23:4:12 2007-5-14 23:6:27 0x01 0x00 0x38 0xff 0x01
(1) When the service is transiently interrupted, the XCE in slot 59 work as the active cross-connect board of the extended subrack. Forced switch the active XCE to the XCE in slot 60, and the service recovers.
(2) Change the XCE in slot 59, and the problem is solved.
LP_SIZE_ERR is an alarm about wrong TU specifications. This alarm is reported when a board receives the TU service whose mapping structure is different from that set in the board. When this alarm occurs, the lower order services are interrupted, accompanied by TU_AIS alarm. LP_SIZE_ERR is reported when bits 5 and 6 of the pointer of the service of VC12 are not "1 0", that is, the pointer is deviated. Deviation of the pointer is most probably caused by the clock.
As something is wrong with the service of the extended subrack instead of the main subrack, the XCE board may be faulty. Actually, the XCE board has no clock module, and it is the main subrack that assigns a 38M clock to the board. The clock combines with the crystal oscillator on the XCE board to provide 8K and 2K frame headers to the service boards on the extended subrack. When the service boards lose clock signal, the TU_AIS, TU_LOP, and LP_SIZE_ERR alarms -are generated. Moreover, as the service on the main subrack has no problem, something must be wrong with the crystal oscillator of the XCE board.
Crystal oscillator failure seldom occurs. When the fault occurs, the only method to solve the problem is to replace the board. Currently, there is no prevention measures.