Operating System - OpenVMS
cancel
Showing results for 
Search instead for 
Did you mean: 

Alpha ES45 Model 2 detecting failed power supplies ?

 
SOLVED
Go to solution
Thomas Ritter
Respected Contributor

Alpha ES45 Model 2 detecting failed power supplies ?

We have several ES45 each with three power supplies. Is there a way to detect a power supply failure other than physically looking at the front panel LED display ?

4 REPLIES 4
Volker Halle
Honored Contributor

Re: Alpha ES45 Model 2 detecting failed power supplies ?

Thomas,

WBEM or SMH should be able to display this kind of information.

Or try RMC, although I'm not 100% sure, if you can issue RMC commands on the console, while OpenVMS is up and running.

Volker.
Duncan Morris
Honored Contributor

Re: Alpha ES45 Model 2 detecting failed power supplies ?

Thomas,

you can issue the RMC commands while the system is up and running (see attachment).

Duncan
Richard W Hunt
Valued Contributor
Solution

Re: Alpha ES45 Model 2 detecting failed power supplies ?

If you wish to monitor these conditions in another way than the excellent suggestions offered by other forum members, might I suggest you look into the following:

F$GETSYI( "FAN_VECTOR" )
F$GETSYI( "POWER_VECTOR" )
F$GETSYI( "THERMAL_VECTOR" )
F$GETSYM( "TEMPERATURE_VECTOR" )

These give you back a set of two-character codes, one per sensor. The vector might be similar to this:

FFFFFFFFFFFFFFFFFFFFFFFF01010101010101

(Might have more FF pairs, the strings are usually 32 characters long.) Reading RIGHT to LEFT (!!!), these are the flags that show you the states of every environmental and internal sensor your system can see. Consult the Owner's Guide for your particular machine model to see where the sensors are located.

You can write a program or script (if that is your "thing") to periodicially check the four vectors. For the power supplies, fans, and thermal vectors, you have three states, which are FF (no sensor), 01 (good), or 00 (bad). These states ARE dynamic as far as I have seen in the few cases where I was able to test them. (For instance, when we had some down-time, I unplugged a power supply to validate that it showed up as "bad" on the vector.)

The temperature vector is a little different than the other three vectors. It shows you the temperature in degrees C for that sensor - in HEXADECIMAL. If you have an FF for the THERMAL_VECTOR slot or a 00 (failed sensor), the corresponding TEMPERATURE_VECTOR slot has no meaning. Otherwise, the temperature value has meaning.

Example: My temperature vector reads, in part,

..."FFFF1516171617"...

The "16" is HEX for 22 decimal degr. C or 71.6 degr. F. The "17" is 23 degr. C or 73.4 degr. F

I cannot use some of the normal monitoring tools because of our Dept. of Defense setup. (That's WAY too long a story for this forum.) Therefore, I had to "roll my own" for monitoring these sensors.

I wrote a little hourly script to sample those vectors looking for any of them to change, and to retain logs of the changes. I can tell before the engineer's message gets broadcasted around our site if the A/C is getting ready to die.
Sr. Systems Janitor
John Gillings
Honored Contributor

Re: Alpha ES45 Model 2 detecting failed power supplies ?

Thomas,

Richards suggestion of using the F$GETSYI and F$GETSYM items may work on your ES45s, but be aware that they're not implemented on all platforms. Indeed, they only work on a small handful of late model Alpha systems.

Make sure you do some sanity testing (such as Richard described) before depending on your results.
A crucible of informative mistakes