Are RSS and chimney offload stable in R2 with HP NC550SFP?
Hi experts, I found intermittent packet loss recurred in our SAP system recently(at least twice since 2/10). We fixed this issue last year by setting up edge port for network teaming in HP DL980 - NC550SFP. However, after restarting Windows in Chinese New Year, the problem recurred recently. I am afraid if there is any network settings, such as chimney offload, RSS, NIC drivers or teaming, affects this. I called MS CSS before and they always recommended me to disable chimney offload and RSS. Ho do I solve intermittent packet loss issue in Windows system? wireshark? Or just disable every additionl function, such as RSS, chimney offload, network teaming and then trials and errors? Becuase our SAP system is chatty, I don't want to disable everything like CSS suggested. I am afriad this will impact network and system performance a lot.
February 15th, 2012 5:22am

Hi, Thanks for posting here. > I called MS CSS before and they always recommended me to disable chimney offload and RSS. Ho do I solve intermittent packet loss issue in Windows system? wireshark? Or just disable every additionl function, such as RSS, chimney offload, network teaming and then trials and errors? Since we cant determine the root cause right now, so by capturing the traffic may will help to narrow down and troubleshoot effectively. If you are concerning about these Windows built-in features then please refer to the suggestions that published by Tod Edwards and get the best practices recommendations for using them. "Give Microsofts Scalable Networking Pack Another Look" article in Windows IT Pro Magazine http://blogs.technet.com/b/wsnetdoc/archive/2011/10/31/quot-give-microsoft-s-scalable-networking-pack-another-look-quot-article-in-windows-it-pro-magazine.aspx Thanks. Tiger Li Tiger Li TechNet Community Support
Free Windows Admin Tool Kit Click here and download it now
February 16th, 2012 1:50am

Hi Tiger, Thanks for your info. Yes, I can't identify the root cause right now. I have read the link and I followed everything he mentioned. Only one hotfix which causes error code(memory dump) but I think it will solve my problem right now. Anyway, I ran wireshark and I found intermittent packet loss affected every host when it happened. SQL1 to SQL2, two Nexus, 4 Nortel, inside the blade servers...just everywhere in the boradcast domain. I have spent more than 1 year and can't identify the root cause(by the way, sometimes it happened only once per month).
February 16th, 2012 2:57am

Hi, Thanks for update. Since you have engaged with CSS, So have they also given further suggestions or maybe we can show the captured results to them for further investigation. Thanks. Tiger LiTiger Li TechNet Community Support
Free Windows Admin Tool Kit Click here and download it now
February 17th, 2012 3:26am

Hi All, The cause is 32GB disabled by BIOS, FYI.
March 5th, 2012 1:59am

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics