Single Bit Errors and a patch?

[Rick Garland]

Question:

A system is generated SBEs. They are become more frequent. Somebody told me there was a patch for this?

This is new to me and a search did not produce anything. Anybody here of this?

Thanks

[Mel Burslan · Accepted solution]

Rick,

Personally never heard such a thing and I would not expect it to be the solution to SBE problems. More frequent but correctable single bit memory errors encourages me to replace the bad memory module as quickly as possible, since this might be the precursor to a total failure of the module. In my opinion again, patching to prevent seeing these errors is like turning your head to the other side when you notice the problems. But again, what you might be looking for can be something totally else than my understanding.

[spex]

Rick,

In my experience, whenever EMS has reported a SBE, RAM was going bad.

PCS

[Jaime Bolanos Rojas.]

Rick,

I believe those are the patches that you are looking for:

PHSS_29343 and 29344

Regards,

jaime

[A. Clay Stephenson]

There were some patches for STM dealing with single bit errors but the patches generally revolved around changes to the descriptions of the problem and changes to the actions you should take. In any event, if you are seeing more than about 1 SBE / week (and that is high) and especially if the errors are confined to a single DIMM then I would have it replaced. Single bit errors can be caused by background radiation including cosmic rays so a few are unavoidable but these should be very rare events.

[Torsten.]

Hi Rick,

the only patch-related thing is a current version of diagnostic.

But I would recommend to open a service call and let HP investigate the errors.

A single error may cause a lot of error messages.

[James R. Ferguson]

Hi Rick:

You can try simply cleaning the DIMM contacts with an alcohol wipe and reseating them. Over time, microscopic amounts of dirt and oxidation can lead to errors that don't really represent faulty memory.

Regards!

...JRF...

[IT_2007]

If you keep on getting SBE on same memory module, it is time to replace.

[Torsten.]

Rick,

let me try to explain it. If the diagnostic is detecting an error, it is trying to test the memory and wants to disable it if needed (only a small 4k block). If the region is in use, it cannot be disabled and the error messages will continue until the memory area is not in use any longer by an application.
A reboot will solve this situation. This is the idea about it, more or less.
In addition, most servers are using ECC and chip spare to work around memory errors.
See this document as an example for chip spare (locate the "Chip Spare Functionality" section):

http://docs.hp.com/en/rx1600_OpMaint/apas02.html?jumpid=reg_R1002_USEN

Anyway, you should give HP a call and let them investigate this.

[Rick Garland]

What I thought.

Will reboot, clean, reseat, and keep an eye on the system.

Thanks!