V2250 Fault light blinking

[JDL_2]

Does anybody know how to stop the fault light bar of a VClass? I thought maybe something like clearing a log much like going into GSP on an N and L class server. I found out that there was a LPMC parity error and then a root disk had a hickup. I also had reports on cstm that a path leading to a fibre mux is logging errors. I power cycled the box and the jamaicas where the root disk in question is, and I went ahead and power cycled the mux's as well. The box came up okay and all is working but the light bar still flashes. I issued an pce_util -b on the test console as sppuser and it did not help. Thanks in advance.

[John Poff]

Hi,

If your pce_util command didn't work, you might still have a problem in the box. The light bar will also flash if there is an environmental problem, such as high temperature. Are all of your fans working on the system?

JP

[JDL_2]

Hi John,

Yes all the fans work in the box. There are also 2 others next to it and all 3 are only about 1 1/2 feet apart. The other 2 are okay. Where else can I go to find out if there's anyhing wrong with the box?

[John Poff]

Hi,

You might need to call HP then if your light bar is still flashing. I used to have some docs about the V-class boxes but I can't find them. They are probably stuck away in a box somewhere.

Your post reminds me about our V-class boxes, which are coming to an end. We got our first V-class box in May of 1998, and we had three at one point. We've replaced them with rp8400 boxes, and the last V is going out the door Saturday. I'm gonna miss the big old lugs. A V-class box may not be the fastest thing going anymore [not even close], but they sure look like they should be fast. I always loved powering one up and how all the fans start going, and all the noise the thing makes. You really had a sense of firing up some serious hardware when they start up.

JP

[JDL_2]

Yeah John, I know what u mean. Ours are ready to go really soon but I'm trying to still maintain it until we power them off permanently. These have served us since 1998 as well and it is just that time. They are going early next year and I have to move her to a new home, rp7400 and her data moved over to newer disks. I need her up until the migration date. Unfortunately we do not have support for this box anymore because she's only historical data but the company needs her running for a couple more years. Thanks again for your input John. I'd like to here more from others who may have any ideas. Thanks again.

[Patrick Wessel]

Hi,
did you ever read the error status of the box? If not, post the output of pce_util and we will see what the system tries to tell us.
In the pce_util output is a field called LEDS. You will find the meaning of this code in the manual page of leds (man leds)
Certain error conditions prevent to switch of the error state of the light bar.

[JDL_2]

Hi Patrick,

Here is the ouput from pce_util -b. I tried it a couple of times and all that changes is that there is a "B" after 0x71. I did a man on leds but I could not make anything of it. Thanks for your help.

pce_util: new node list: 0
Complex name is Default_complex
Node Clocks LEDS @C U SHPT Supply 1 Supply 2 Supply 3 Supply 4
---- ------ ------------- ---- ------ -------- -------- -------- --------
0 Normal ATTN 0x71 21 1 0000 Nominal Nominal Nominal Nominal

[Eugeny Brychkov]

0x71 is one of the power supplies (do nto remember which ne). Check that all are connected, turned on and functioning. BTW, there's a switch on one of boards which controls power redundancy. Check it
Eugeny

[Patrick Wessel · Accepted solution]

Hi,
As Eugeny already explained failed one of your 48V power supplies. When you enter the following command on your system support processor:
man leds | grep 71
is the output:
71 NPSUR failure (When viewed from Node front)

(71 is the error code in your LED field of the pce_util output)

NPS: node power supply
UR: upper right

You will see max. 4 power supplies in the system when you open up the large panel on the front. The upper right reported an error
Power failures need to be reseted before you can switch the light bar to solid. This reset is usually a power cycle. This prevents the system from acting like a yo-yo in case of a serious power problem (powering up, report a serious power problem, going down, power back on, reporting the problem, going down...)

[JDL_2]

Hope everybody had a nice holiday! Patrick, thanks for explaining the output of the pce_util mean. We power cycle'd the box just before I issued the pce_util command and right before I posted this thread. It did not reset the blinking led. Do you think a couple more power cycles will do? Eugeny, thanks for the tip regarding the switch on the board. Will this disable the alert signal from the bad power supply?

[Eugeny Brychkov]

This little switch is called 'redundancy switch'. It is on Core utility board (the one with leds in the center of V-class looking from the front). If it's off (down position) then ALL installed power supplies should be operational otherwise V-class will not switch on. If it's in on position (up) then n-1 of n power supplies should be on and functional. In case when this switch is off (all PSUs should be on) AND one power supply dies V-class continues to operate but shows error. If in this situation you will power cycle V-class it will not turn on. As I understand default position of this switch os ON
Eugeny

[Patrick Wessel]

JDL,
Hope you had a nice Holliday, I had the add some vacation, this is the reason why I answer with a little delay.
Based on your description of the situation, I???m afraid that we deal with a bad power supply (should be the upper right). Every power supply has it???s own DC_OK signal, that is monitored by the Monitoring Utility Chip (MUC) on the Core Utility Board (CUB). If a power supply is installed and does not report DC_OK it will be marked bad by the CUB.
How many Node Power Supplies are in your V-Class? Have you checked the status LEDs on the NSPs? The yellow LED indicates that AC power is applied to the NSP and the green LED indicates that the NSP is operating and providing 48V dc. If the yellow LED is of, verify that the NPS is switched on by the power switch on the NPS

[JDL_2]

Sorry for the delayed replies guys. I've had a lot of production issues during the last couple of weeks.

Eugene, I am assuming that the redundacy switch is on because all leds on the the PSU's are on. I have not physically check it. Thanks for that info. And to Patrick's question, there are a total of 4 NPS's and all except for the one on the top right does not have the green led on. At this point our migration date is right around the corner and a have a pretty good set of good backups of the machine. In any case I'll bring it up on another machine. I'm just hoping it'll hold on for another month. Thanks for all your input. I've learned quite a bit on this thread.

[Patrick Wessel]

The situation sounds not to bad. 3 power supplies are sufficient for a V22x0 the forth is for redundancy. So, you don???t have any redundant power supply anymore and the bad one is marked very clearly. It could be worth???

[pj_11]

Hello JDL,

Try the follwoing command i have used it and it has worked for me.As to PSU issue you may
need to get this sorted out.The following command pce_util -n 0 -r 1


Kind Regards,

PJ.