ML350 G5 Rebooting every two weeks

[Roddi Smith]

I have an ML350 G5 running Win2K3 Server. The system restarts itself every two weeks to the day at random times, i.e. sometimes 7am, sometimes noon. The server was in service from Aug 2006 - Jan 2007 with no problems.

Beginning 2/10 the server restarts itself every two weeks at different times of the day.

I initially thought it was a Windows update issue and turned off automatic updates. When it happened again I ran hardware diagnostics and found nothing. On 3/28 one of the power supplies failed and was replaced. This failure generated the usual warning in the event log and did not cause a restart or a shutdown.

Two weeks to the day after I replaced the power supply the server restarted itself again.
At HPs suggestion I updated the BIOS and Firmware and the server still restarts itself every two weeks. I have determined that the restart happens two weeks from the last shutdown or restart, planned or not. I am working around the problem for now by deliberately restarting the server at times when it will not affect my production staff but I can't keep this up forever.

Any assistance is appreciated.

[Dave Houseman]

Hi,

Will you please tell me if you have resolved this issue ? I have the identical problem and have spent many hours attempting to identify the cause.

Please reply.

Thanks.

[Alan Steel]

Hello Roddi and Dave,

I wish I had an answer for you, but unfortunately I am experiencing the same problem. Did either of you ever find a fix?

Regards,

Alan

[Dave Houseman]

I recently learned that some APC UPS units perform a self-test every 14 days. If a battery is bad, or going bad, the UPS may restart.

[Alan Steel]

Hello Dave,

Thanks for checking in.
I just tripped over the same information!
I'm not sure if its relevant to me, even though I am using an APC UPS.
I say this because:
Nothing else connected to the UPS is having this problem;
My reboots are happening on the 23rd and 25th of each month (January and February so far), at different times of day;
The 23rd "happens" to be exactly 2 weeks after I installed this server;
The UPS it is connected to is brand new, with no apparent problems.
That being said, I may just try a different UPS for troubleshooting purposes.
I'm thinking its probably power-related or the SATA RAID susbsystem doing some process on a schedule that takes the system down.
Other than that I'm completely stumped: there are no errors in any of the logs, and none of my other servers are having problems around the same times, I have a very generic installation -- one MS SQL app, no 3rd party hardware.

Thanks again,

AGS

[Dave Houseman]

We had the same issue ... brand new APC UPS ... no events logged on server ... no other problems with hardware.

However, we tracked the reboot to be exactly every 14 days from the last reboot, regardless of the reboot reason.

We have recently attached another new APC UPS to our server. Since we have redundant power supplies, we've plugged one power supply into a surge protected outlet and the other power supply into the UPS.

If our assumption about a bad UPS is correct, and we are unlucky enough to have a second faulty UPS, we should log an event in 14 days without a server reboot.

I'll keep you informed.

[Dave Houseman]

The problem is actually an APC UPS incompatibility issue. It seems that at least the APC Back-ups model "XS - 1500" causes the server to reboot.

We installed a brand new "XS - 1500" onto the server. Since we have redundant power supplies, we plugged one into the APC UPS and the other into a power strip.

14 days since the last reboot and we got a power supply failure written to the event log.

It's the UPS.

We will replace with a different brand.

[RaMpaNTe]

The APC 1500 UPS is well known for causing rebooting and power down issues in several G5 models. We have fixed tons of issues that even after replacing parts nothing seemed to work.

Everything was resolved once the customers decided to stop using the ACP1500 UPS.

I know the VA are less but it has to do with the frequency. Try an HP 1500 you'll see that everything will be fine.

Thanks,
\RaMpaNTe