Users may sometimes see error logs indicating the parity error "PE Chip::FATAL ERROR!! from PE5[5]: DLU: elp prot0 detected_rsb1 error" in DLU on PTX Series routers.

This article explains that these are parity errors that are considered transient and that should clear on their own. In case the error does not clear on its own, it is suggested that the FPC be physically re-seated and the chip reset.

The following FATAL error logs may be seen on PTX Series routers with PE-chip based Flexible PIC Concentrators (FPCs), which subsequently disables the Packet Forwarding Engine (PFE). In fact, the PFE will not come up even after the error is cleared.

​Mar 30 12:11:41.771  labbox fpc2 Cmerror Op Set: PE Chip::FATAL ERROR!! from PE5[5]: DLU: elp prot0 detected_rsb1 error 

Mar 30 12:11:41.815  labbox fpc2 ISSUING ifd-down for et-2/0/21 (580) on pfe-5
Mar 30 12:11:41.817  labbox fpc2 ISSUING ifd-down for et-2/0/23:0 (582) on pfe-5
Mar 30 12:11:41.817  labbox fpc2 ISSUING ifd-down for et-2/0/23:1 (583) on pfe-5
Mar 30 12:11:41.818  labbox fpc2 ISSUING ifd-down for et-2/0/23:2 (584) on pfe-5
Mar 30 12:11:41.819  labbox fpc2 ISSUING ifd-down for et-2/0/23:3 (585) on pfe-5
Mar 30 12:11:41.819  labbox fpc2 ISSUING ifd-down for et-2/0/25:0 (590) on pfe-5
Mar 30 12:11:41.820  labbox fpc2 ISSUING ifd-down for et-2/0/25:1 (591) on pfe-5
Mar 30 12:11:41.821  labbox fpc2 ISSUING ifd-down for et-2/0/25:2 (592) on pfe-5
Mar 30 12:11:41.821  labbox fpc2 ISSUING ifd-down for et-2/0/25:3 (593) on pfe-5
Mar 30 12:11:41.822  labbox fpc2 ISSUING ifd-down for et-2/0/29:0 (598) on pfe-5
Mar 30 12:11:42.661  labbox fpc2 ISSUING ifd-down for et-2/0/29:2 (600) on pfe-5
Mar 30 12:11:43.290  labbox fpc2 ISSUING ifd-down for et-2/0/29:3 (601) on pfe-5
Mar 30 12:11:43.747  labbox fpc2 pechip_set_pfe_disabled:976: PECHIP[5]: PFE Marked Disabled !!!
Mar 30 12:11:43.750  labbox fpc2 pechip_set_pfe_disabled:983: PECHIP[5]: Deregisterd TOE PIO Vectors !!!
Mar 30 12:11:43.750  labbox fpc2 JPRDS_FDB:ERR:jprds_fdb_set_pfe_disabled(),7317: pfe 5 marked disabled  

Mar 30 12:13:03.294  labbox fpc2 Cmerror Op Clear: PE Chip::FATAL ERROR!! from PE5[5]: DLU: elp prot0 clear detected_rsb1 error  

This log message indicates that a parity error was observed in the odd blocks of the Result state buffer (rsb1), which caused the PFE where the error was seen to be disabled in order to ensure that the other PFEs were not affected.

As you can see in the logs above, the error clears on its own even though the PFE remains disabled because this is a fatal error and needs a hardware reset to clear it. You can confirm the same from the below output of the FPC.

1. Log in to the affected FPC, in this case FPC2:

> start shell pfe network fpc2

2. Run the following command to confirm the same:

 bringup jspec read pechip[0] register dlu elp prot0 log_address
 bringup jspec read pechip[1] register dlu elp prot0 log_address
 bringup jspec read pechip[2] register dlu elp prot0 log_address
 bringup jspec read pechip[3] register dlu elp prot0 log_address
 bringup jspec read pechip[4] register dlu elp prot0 log_address
 bringup jspec read pechip[5] register dlu elp prot0 log_address

Example

FPC2(labbox vty)# ...ip[0] register dlu elp prot0 log_address
0x09958010  pe.dlu.elp.prot0.log_address       00000000
FPC2(labbox vty)# ...hip[1] register dlu elp prot0 log_address
0x09958010  pe.dlu.elp.prot0.log_address       00000000
FPC2(labbox vty)# ...hip[2] register dlu elp prot0 log_address
0x09958010  pe.dlu.elp.prot0.log_address       00000000
FPC2(labbox vty)# ...hip[3] register dlu elp prot0 log_address
0x09958010  pe.dlu.elp.prot0.log_address       00000000
FPC2(labbox vty)# ...hip[4] register dlu elp prot0 log_address
0x09958010  pe.dlu.elp.prot0.log_address       00000000
FPC2(labbox vty)# ...hip[5] register dlu elp prot0 log_address
0x09958010  pe.dlu.elp.prot0.log_address       00000045   >>> log_address showing a non-zero value for the affected PFE, in this case pechip5

Note: The bringup jspec command is used to read the register values only and does not have any impact on production. It can be run on a production device.

To clear the parity error and reset the chip, you would need to physically re-seat the FPC and reset the chip.

Typically, the above error is considered to be a transient hardware error that should not recur in the same location. However, if you see the error again in the same location as mentioned in the log_address output above, you would be required to RMA the FPC with help from Support .