hardware problem with this HBA?

[stephen peng]

Guys,
I've got problem with one fibre link to the eva4000:
ioscan -fcnCfc
Class I H/W Path Driver S/W State H/W Type Description
===================================================================
fc 0 0/0/10/1/0 td CLAIMED INTERFACE HP Tachyon XL2 Fibre Channel Mass Storage Adapter
/dev/td0
fc 1 1/0/4/1/0 td CLAIMED INTERFACE HP Tachyon XL2 Fibre Channel Mass Storage Adapter
/dev/td1
and fcmsutil told:
fcmsutil /dev/td0

Vendor ID is = 0x00103c
Device ID is = 0x001029
XL2 Chip Revision No is = 2.3
PCI Sub-system Vendor ID is = 0x00103c
PCI Sub-system ID is = 0x00128c
Topology = PTTOPT_FABRIC
Link Speed = 2Gb
Local N_Port_id is = 0x010400
N_Port Node World Wide Name = 0x50060b000030b72f
N_Port Port World Wide Name = 0x50060b000030b72e
Driver state = ONLINE
Hardware Path is = 0/0/10/1/0
Number of Assisted IOs = 1978830891
Number of Active Login Sessions = 2
Dino Present on Card = NO
Maximum Frame Size = 2048
Driver Version = @(#) libtd.a HP Fibre Channel Tachyon TL/TS/XL2 Driver B.11.11.12 PATCH_11.11 (PHSS_31326) /ux/kern/kisu/TL/src/common/wsio/td_glue.c: Sep 5 2005, 10:14:40

eva4k74[/]#fcmsutil /dev/td1

Vendor ID is = 0x00103c
Device ID is = 0x001029
XL2 Chip Revision No is = 2.3
PCI Sub-system Vendor ID is = 0x00103c
PCI Sub-system ID is = 0x00128c
Previous Topology = PTTOPT_FABRIC
Link Speed = UNINITIALIZED
Local N_Port_id is = 0x010400
N_Port Node World Wide Name = 0x50060b000030321f
N_Port Port World Wide Name = 0x50060b000030321e
Driver state = AWAITING_LINK_UP
Hardware Path is = 1/0/4/1/0
Number of Assisted IOs = 787590227
Number of Active Login Sessions = 0
Dino Present on Card = NO
Maximum Frame Size = 2048
Driver Version = @(#) libtd.a HP Fibre Channel Tachyon TL/TS/XL2 Driver B.11.11.12 PATCH_11.11 (PHSS_31326) /ux/kern/kisu/TL/src/common/wsio/td_glue.c: Sep 5 2005, 10:14:40
now LUNs status from 1/0/4/1/0 are all NO_HW, left light of that HBA is on, and right light, off; from SAN switch's switchshow, the port this HBA connected is in the status"NO_LIGHT"
can I make sure that this HBA encounters hardware problem and need to be replaced?

regards
Stephen

[Jose Mosquera]

Hi,

Driver state = AWAITING_LINK_UP means that your card is dead.

Rgds.

[Jose Mosquera · Accepted solution]

Hi again,

You must discard first any issue on cable, connectors (server/SAN switch) or switch ports.

Rgds.

[Torsten.]

AWAITING_LINK_UP means there is no link between the card and the switch.


This could be

- bad HBA
- bad cable
- bad switchport

[Jorge Pons]

Hi

Its probably that you have any problem in cable, or switch port. check it.
When a hba start, both (hba and switch) talk about arbitration, and in this case, not ocurr.

regards, Jorge

[Manix]

Check the the SAN-switch port settings .Are all lights on the FC-connectors green ?

This is a common error as said earlier due to
switch /HBA /cable.

[stephen peng]

following is recent dmesg of this server:
ct_query failed. hw_path = 1/0/4/1/0
ct_query failed. hw_path = 1/0/4/1/0
ct_query failed. hw_path = 1/0/4/1/0.
I swap that cable to another port of the switch, no green light on both port at both cases.
It is time-consuming to get a new HBA or a new cable, so I do want to establish which part of the link failed.
If the HBA died when the os was running, would there be log somewhere?


regards
Stephen

[Jorge Pons]

Hi

what is the output of:
# ioscan -funC fc
# ioscan -fnC fc

Regards, Jorge

[Torsten.]

You will likely see anything like a "link lost"; but depending on what is bad now, the server or the HBA itself is not really aware of it.

You said there still a light comes out of the HBA (be very careful, it is a laser and can damage your eyes!); is this light also visible at the switch port and the cable connector?

[Jorge Pons]

Hi

Can you test in another switch port?

[Jorge Pons]

Sorry
I didnt read that you test in another port

[stephen peng]

Torsten,
About HBA's light, I meant that two status light on that HBA(one for "ACT" and one for "LNK"?), not from inside the cable, "ACT" was on and "LNK" was off; from the good one(HBA", "ACT" was blinking and "LNK" was on.

Jorge,
ioscan -fnCfc was at the first beginning of this thread, and I changed the cable port to another on that switch, it remained the same.

regards
Stephen

[Torsten.]

>> I changed the cable port to another on that switch, it remained the same.


What remained the same?
The fcmsutil output?


An easy mothod for testing is to plug in a loopback connector, if you have one.


However, if light comes out of the transceiver (be careful, I told you!), the HBA is likely OK. If no light comes out of the cable, the cable is likely bad. Same for the switch and the other direction of the cable.

[Jorge Pons]

Hi

Like said Torsten, you can "test" the cable.
connect to hba and see at other connectors (carefully) if light is there. then change position of cable to see the other one.

Regards, Jorge

[stephen peng]

Torsten,
>>What remained the same?
The fcmsutil output?

The HBA's status LED and the new port's status LED remained the same.

>>light comes out of the transceiver .
what dose "transceiver" mean here?

>>However, if light comes out of the transceiver (be careful, I told you!), the HBA is likely OK. If no light comes out of the cable, the cable is likely bad. Same for the switch and the other direction of the cable.
please explain in more detailed,thanks!

I'am not an expert to "test" a cable.

regards
Stephen

[Torsten.]

The light comes out where the cable plugs in ...

[Jorge Pons]

HI

to test the cable is very simple (to see if its broken).
The hba (or switch port) has two "holes" one is for transmit and the other one is for receive. It means, one emited red light, and the other one receive the light that emited other device.
If you see directly this holes (hba or switch port) you can see this light, but carefully because if you see for long time, you can damage (one second, dont worry).
The cable, is made of crystal (or something laike this) and it reflexion the light from end to end, and its possible to see the light also.

for this, if you plug cable into hba and see the end of cable, you must see light. If you see light, this part of cable its ok. then you change the position of cable to see the other one.

Regards