ASR 9000 Power Supply Failure: PSM Troubleshooting

A power supply module failure on an ASR 9000 rarely announces itself as a power problem. The first symptom is usually a line card that refuses to boot or a driver initialisation delay, because the shelf manager will not allow cards to power up until the chassis power allocation is known. This runbook follows the actual Cisco troubleshooting path: confirm the platform state, read per-module input and output power, inspect the environment, check known bugs, then either reseat the module or collect PID and serial evidence for an RMA. It also covers the V1 DC power module defect in which supplies disappear from inventory entirely.

Step 1: identify the failed module

RP/0/RSP0/CPU0:ASR-9901-A# show platform

Node          Type                 State          Config state
0/RSP0/CPU0   ASR9901-RP(Active)   IOS XR RUN     NSHUT
0/FT0         ASR-9901-FAN         OPERATIONAL    NSHUT
0/0/CPU0      ASR9901-LC           IOS XR RUN     NSHUT
0/PT0         A9K-AC-PEM           OPERATIONAL    NSHUT

If all power trays report OPERATIONAL, move to the per-module view. That is where a single failed supply inside a healthy tray shows up:

sysadmin-vm:0_RSP0# show environment power

Total output power capacity (N + 1) : 1600W + 0W
Total output power required          : 930W

Power Supply   ------Input----  ------Output---  Status
Module  Type   Volts  Amps      Volts  Amps
0/PT0-PM0  1k6W-AC  216.0  2.4  12.0  38.1   OK
0/PT0-PM1  1k6W-AC    0.0  0.0   0.0   0.0   FAILED or NO PWR

A module reporting zero input and zero output is either unpowered or dead. A module reporting input but no output is in a power limit or overcurrent state.

Step 2: environmental and physical checks

• Verify ambient temperature and airflow; a supply that is overheated derates and eventually faults.
• Confirm the source is within specification - 200 to 240 VAC for AC modules, -40 to -72 VDC for DC.
• Check retention clips and terminal studs. Loose input connections produce the classic blinking-green input LED.
• Read the LEDs: solid green input, solid green output, red fault. Blinking output means power limit or overcurrent.

RP/0/RSP0/CPU0:ASR-9901-A(admin)# show environment
RP/0/RSP0/CPU0:ASR-9901-A# show logging | include Power

Step 3: known defects to rule out

V1 DC power supplies generate I2C errors when they have no input voltage. If a whole tray of DC supplies vanished from admin show platform after a power-feed test, the modules were marked failed by the power manager rather than actually dying. The documented recovery is a process restart, not an RMA:

sysadmin-vm:0_RSP0# process restart pwrmon
sysadmin-vm:0_RSP0# process restart shelfmgr

Order matters: pwrmon first, then shelmgr. Otherwise the shelf manager can refuse card power allocation on the next boot.

Step 4: reseat or replace

Perform a jack-out/jack-in on the suspect module, then re-read admin show environment power. If the fault follows the module into another slot the module is faulty; if it stays with the slot, the power tray or PEM backplane connector is the problem. Collect evidence before opening a TAC case:

RP/0/RSP0/CPU0:ASR-9901-A# show inventory location 0/PT0
NAME: "0/PT0-PM1", DESCR: "1600W AC Power Module"
PID: A9K-1600W-AC, VID: V01, SN: POGXXXXXXX

Preventing repeat incidents

Keep N+1 headroom above the calculated chassis requirement, monitor the input-versus-output delta as a leading indicator of supply degradation, and record insertion counts. For broader chassis-level diagnostics around module discovery and database state, see ASR 9000 IM and SYSDB debugging and IOS XR show tech-support collection.

原文链接：https://www.cisco.com/c/en/us/support/docs/routers/asr-9000-series-aggregation-services-routers/225782-troubleshoot-power-supply-module.html