Problem detected

[SeanAIX430]

OK my server is giving me this message:

Periodic diagnositcs for sysplanar0 has detected a problem.
The Service request number is 805-805

I haven't found any answers in my online searches, and help would be welcome, thanks.

 

[aixqueen]

Service Request Number List for Multiple Bus Systems


Replace FRU parts in the order by which the "Failing Function Codes" are
listed.

Service Failing Description and Action
Request Function (Unless otherwise indicated use Chapter 12.
Number Codes "MAP 0210: General Problem Resolution".)
SRN
Src.
805-805 G 152 Description: System shutdown due to loss
of AC power.
Action: Check the voltage range switch, if
present, on the power supply and ensure
that it is set to match the AC input
voltage. If the range switch is set
correctly or is not present refer to "MAP
1520: Power" in the Service Guide for your
system.

 

[SeanAIX430]

Think the fact that I had to do a hard reboot is the actual issue? Would a clean shutdown and restart clear the error message?

 

[aixqueen]

I would check the switch for power on the back first........
More than one occurance in the log?

It could be a hard reboot............you could clear the errpt with the errrclear 0
or send to a file errpt -a >>/tmp/mylog
and then clear it........and see if it reoccurs....If it doesn't dont worry,
if it does...you will need hardware or check the switch or check the ac/dc switch
for the power supply or the plug in on the back.

 

[samizdat]

We replaced the I/O in two of our M80s. Obviously, we had to remove power and they both came back up with this message. The CE told me the reason was the removal of power. Just to be sure, I cleared the errlog and had the CE hang around while I rebooted, both servers came up normally.

 

[gheist]

errclear 0 helps here (and with any error happening every day like rmt0 screming for cleaning or ses0 for cabling) - and diag to adjust time when regular diags run / there are other ways to fast shutdown eg # killall ; halt on sys console. If you got this prob more often - it might be probl with your firmware (pls practice on cd0 before sysplanar0)

 

[Yegolev]

The machine can tell whether you poke the power button or yank the cable. To keep diagela from bothering you, you can either clear the error log (errclear 0) or remove the resource from the diagela list from within the diag program.

Note that not everything in the error log is an error, some are merely informational.

 

[SeanAIX430]

This reoccuring problem that is pointing to a failiing power source makes me nervous. So I want to order another power supply for this machine in case the present one goes out. Problem is I can't find the part number, hopefully someone knows and I won't have to open the machine up. It is an IBM 7024 E20 rs/6000 box. I've been searching for the part number but have been unlucky so far. Anyone know what the part number is?

Thanks

 

[Yegolev]

FYI, the E20 has very bad airflow. Make sure it is not an overheating problem, although that should appear as an environmental problem. The E20 may not have the proper sensors to register that, however. No need to replace the psu if it is a temp problem, you know... but it is probably the psu.

 

[avadh]

Hello,
If you are in USA and have IBM hardware service contract then call 1-800-IBM-SERV - they will either send you hardware support(if you go to hardware support they will take service request number) order the part and send the call to IBM CE who will give you a call to comeout and change the part.

IBM also has a service call Service director and new GUI based Service AGENT once it is setup and any hardware failes(SRN=service request number) then it calls IBM to say this part has failed. Part is order and IBM CE calls customer.

 

[avadh]

Hello,
The answer to the problem is this.
Since periodic diagnostics runs everyday at 2-4AM in the morning.

It checks errorlog and if it sees hardware permanent error it runs diag on it and if it generates the SRN then diagela send's mail everyday.

To fix this problem
if you need to save the error log.
errpt > error.txt or
errpt -a > error.txt

then do
errclear 0(zero)

this will clear error log.

Wait till tomorrow to see if you get a e mail from diagela or not.

This will stop diagela from sending you this message.

 

[SeanAIX430]

Yeah I did a clean restart yesterdat to clear the error log and when I got in this morning I got the same error message. So it's recurring, unfortunately. Plus I don't have and service contract with IBM for this machine. I'm just trying to find the part number so I can go online and order the psu from somewhere.

Thanks

 

[gheist]

you have to clean error log with command "errclear 0" - not restart

 

[Yegolev]

Check root's crontab. Look for something related to emailing an admin about power.

 

[avadh]

Hello,
gheist is right - you have to do errclear 0 then do errpt -a | more and if you do not have any error then wait till Friday and then update the forum on results.

 

[aixqueen]

errclear 0 first though to see if it clears...

7024 E20 93H3504 <-- 40H2115 power supply
40H2120 cable with start/stop switch

 

[SeanAIX430]

OK I did errclear 0, and when I came in this morning I didn't have a new error. So that is good news, but since the error report does show several loss of electrical power errors and its plugged into a UPS, the psu might still be the problem. Thanks everyone!

 

[SeanAIX430]

Also I did a search on google with the part numbers that Aixqueen gave me and I didn't come up with any finds! Any ideas where I can get these parts? Thanks again!

 

[aixqueen]

That is the Fru Part number.... maybe a reseller can match the IBM FRU with
the real thing?
Failing Machine FRU Description and Notes
Funct. Type/Model Part
Code Number
152 7024 93H3504 Power Supply

Think the machine is sunset so it would be reseller, remarketer, or refurb?
Marketing withdrawn 7/18/1997

 

[Yegolev]

Must obtain from IBM... or look to buy a E20 from EBay for parts. Check other used RS6k sites is my suggestion. E20 has been out of production for a bit, 1997?

Also might not hurt to make sure the cables are plugged in nicely. Those numbers may be either part numbers or FRU numbers, but are IBM numbers. Good luck.