06-30-2014 07:00 PM
We have a small number of 93z desktops scattered around our school. There seems to be an issue with the ethernet adapters running the Intel I217 chipset. When these machines are in sleep one or more start sending IPV6 "Multicast Listener Report Message v2" packets to the network (observed by Wireshark). Then they begin to snowball and shortly thereafter they are flooding the network, bringing it to its knees.
We had the same issue on some Dell workstations, however, Dell released a firmware update to address this issue with this chipset.
Just wondering if this is an issue Lenovo is aware of and/or has addressed. We've worked around it on Lenovo's machines by disabling all power options on the ethernet adapters but we shouldn't have to do that. If Dell can provide a firmware fix why can't Lenovo (assuming that they haven't).
I know I'm not the only one - https://www.google.com.au/?complete=0#complete=0&q=intel+I217+flood&safe=active
10-20-2014 01:53 AM
I've got exactly the same issue with my Lenovo M93p. The problem is already raise. Please, Would you tell us which is the network driver we have to use to solve the issue ?
10-20-2014 09:47 AM - edited 10-20-2014 09:53 AM
Are you running the Ethernet driver distributed by Lenovo, or the latest driver downloaded from the Intel Download Center (website)?
10-20-2014 02:35 PM - edited 10-20-2014 02:36 PM
I can't tell you which drive to use because Lenovo are terrible with drivers, however, the best way to workaround the issue is to disable the power management settings on the adapter...
We had the same problem on some Dell T1700 workstations, Dell moved very quickly to release an updated BIOS and driver that resolved it. Intel, I believe, have also corrected this in their native drivers but I'm unsure (and unwilling to test) if Lenovo have, I don't believe they had as of a month ago as we reimaged some machines and updated them with latest drivers through the update tool and then we started seeing multicast traffic again.
Lenovo support is rubbish.
10-21-2014 02:29 AM
We use the 12.10.30 from lenovo.
Thank you for you advice I will try to push the configuration about power management (more than 400 computers).
Last thing: are you able to repdroduce easily the issue in lab ?
10-21-2014 02:50 PM
We had it occur on our production network but was able to reproduce it in a small lab (although not to the same level of extremity). You need more than one machine, we had about 12 in a music lab. Once they have been asleep for a while, one of them will send a packet and then so will another and they seem to snowball and you end up with all these packets killing the network. We noticed when we started having intermittent poor ping times to servers, switches and the net, then running Wireshark we found all the packets and narrowed it down to specific machines. These machines all had the same family of NICs. Dell fixed it quick, Lenovo did not.