ITEM: DE2508L
System attached to UPS randomly shuts down.
ENV: AIX 3.2.5 model 7013-570
PROBLEM: system down with a tcp/ip daemon error
ACTION TAKEN: She is rebooting right now and wants to call me back
but wants to know why the system went down?
We can not determine yet; we need to wait for the
system to come back up to look at the error report.
Also, we can look into the problem with TCP daemon.
. System came back up fine.
errpt -a -> tty overrun, processor overload
The system logged 70 of these messages in the error
report.
ACTION PLAN: Research.
TEST CASE: n/a
ACTION TAKEN: Customer has 570 with UPS (600kva). Customer has
system which shuts down randomly. Probably due to
UPS. Error log has tons of TTY_OVERUN errors for
tty1, 13, and 56. This may be unrelated. These
were the only errors in error log.
ACTION PLAN: Research.
Response:
ACTION TAKEN: Customer's UPS is attached to the S2 port of the
system. This corresponds to tty1.
. Every Friday afternoon, the system is shutdown and
powered off. It is then brought back up to do the
system backup and their weekend processing.
. This problem began appearing last Thursday and the
number of system shutdowns has been increasing.
. The network interfaces are then brought down and
the shutdown appears to hang there. It does not
ever get to "halt completed"
. I've asked customer to send me her error report for
today, since the system has already shutdown three
times today. Customer will email this to me.
. Additionally, I recommend that customer try to get
a system dump, since it appears that the system is
hanging during the shutdown.
.ACTION PLAN: Waiting for error report, look for power suspension
errors.
ACTION TAKEN: I've received the error report, there are a lot
of errors, but it only goes back to 11:30 today,
although we generated the report from 00:00 today.
All of the errors in the error log are TTY_OVERRUN
errors, with the exception of an error daemon turned
on entry from the last reboot. There are no power-
related errors prior to the error daemon entry.
.ACTION PLAN: Continue research.
ACTION TAKEN: Customer's system has shutdown again, and she got the
exact wording of the broadcast message:
. broadcast message from unknown on servo
low battery prepare for immediate shutdown
broadcast message from unknown on servo
system is shutting down
The system then starts shutting down and stops putting
status messages on the screen after the TCP daemons are
shut down. Customer does not see a 'halt completed'.
Customer does see the power light go off on the server,
however, so it is turning itself off.
. Customer has noted that the system is shutting down
when someone prints a job. Basically, she hears the
printer start up, then sees these broadcast messages
on the system console.
. Customer is using a Printronics printer. She's not
sure if it is parallel or serial attach.
. Conferenced customer with hardware support to place
a service call. A CE is being dispatched to check
out the power routing between the breaker, UPS and
the system.
.ACTION PLAN: Customer will call me back when the CE gets onsite.
ACTION TAKEN: Customer has turned off the printer, and found
that the system did not shut down when print jobs were
submitted if the printer was turned off. The printer,
a network router and the 570 are all attached to the
same UPS. What was happening was that the drain on the
UPS produced by the printer left it with insufficient
power for the 570, so the UPS shut down the 570.
. Customer is planning to upgrade her UPS and would like
to know what the power draw is for the 570. The power
source loading is 0.43 kVA, and the power requirements
(typical) are 425 watts. Customer will also have to
take her TTY devices into consideration, since the
monitors can draw quite a bit of power. The power
requirements for the monitors are generally printed
by the power cord. As a temporary workaround, I
suggested that customer go ahead and plug her Print-
ronics line printer to a wall outlet with a surge
protector. Line printers are generally OK when they
lose power; the surge protector would protect it from
the sudden unexpected power on.
.ACTION PLAN: Closing item with customer approval.
Support Line: System attached to UPS randomly shuts down. ITEM: DE2508L
Dated: February 1998 Category: N/A
This HTML file was generated 99/06/24~13:30:16
Comments or suggestions?
Contact us