I did this for each pair of DIMM's (but not all possible combinations). I then booted the machine with only one proc and 2 DIMM's at a time. On the first server where this happened, I was asked to pull all DIMM's from the machine and one proc as well. I have 4 other vSphere hosts that have had this issue and had the system boards replaced (eventually). Just for clarity, I have seen the above mentioned line in all of the previous PSOD's. I was stunned at how quickly the turn-around in attitude happened. Let me see if I can get another system board. I mean to say that he went from reluctant to saying, ahh I have seen this before and it has caused lots of problems. He stated that the line containing a reference to "APIC: Lint1 interrupt on pcpu 0" seemed to cinch it as a problem with the system board. Strangely on the most recent (I'm on hold with HP now) the engineer seemed very reluctant to send a replacement system board, but the tune changed dramatically when I sent a screenshot of the PSOD as requested. I have had to be pretty hard-nosed to get the system boards replaced. After that, the servers are much more stable, some have had no issues in MONTHS since mobo replacement, instead of running a max of a few weeks between PSOD. We have seen several issues just as described above and have eventually gotten the system boards replaced. Some running vSphere and some running RHEL 4. He said this usually presents itself as an error with memory on banks 1 to 8 (processor 1), so if you are having random crashes and memory errors involving banks 1 through 8, this could be you issue too.
This is not public, so the HP engineer could not send me the article.but it explains alot. Anyhow, back to the issue.Īpparently there is an 'internal' article about bad D元85 G6 motherboards and they are trying to find the pattern, or batches of these that are affected. I had a HP 'customer advocate' and everything working with me, I am quite the complainer and deal maker (or breaker) sometimes. So I solved a big problem for HP's distribution channel and all they gave me was a $20k refund from the nearly million I spent on a bunch of hardware. I bitched and moaned about lost time and money setting up 5 fully populated c7000 blade systems.
#Hp dl360 g6 memory mac
Apparently during the QC process, they boot them in a computer, and once that happens, the usb key hard codes the mac address and cannot be changed without re-flashing the entire usb key. I was one of HP's first customers to get the HP branded ESXi 3.5 usb keys.I got about 70 and found that there were pre-bound mac addresses on them, and once I started getting most of them online, I had duplicate MAC on my network.I had a hell of a time figuring that one out. Why is it I always seem to find bugs in new products. I asked him if I should changed it to the 'standard' way of largest DIMMs in the first slots.he said just leave it if there is no problem.and he wants me to try that configuration in my current crashed G6. One of my G6's actually has the 4GB DIMMs in the first banks for each processor and has never crashed. I showed them the documentation that said otherwise, and then they changed their tune, although I think this was the norm on other models previously. Now.I have called HP a few times now.and every tech is telling me the opposite, that the SMALLEST DIMMs should go into the slots the farthest from the processor. "DIMMs must be installed in decreasing capacity with the largest DIMMs installed in the banks farthest away from each processor." I have 80 GB total of memory in each of these, eight 8GB DIMMs and four 4GB DIMMSs. I think it might have to do with the order in which my memory was configured. The only things I found were single memory errors for a pair of DIMMs. Screen was blank and I had to hard power down and up the crashed server.very weird. I have had 2 or 3 of these crashing about every month. My most recent crash, the server actually was beeping and showing the banks 1-8 blinking in orange on the front of the server, and the health light blinking red.previous crashes the server(s) was all green, and there were no error lights. I have been having alot of problems with random server crashes with my D元85 G6 servers (I have 4 of them, dual proc six-core), and I believe them to be memory related.
#Hp dl360 g6 memory update
(want to update this thread for potential googlers)
0 kommentar(er)
