SNA DLC traffic stopped passing after upgrade to ESX 3.5

[DHBob]

Looking for any ideas. I have 5 hosts in a VI3 HA cluster. I have a MS SNA gateway (VM guest machine) server using SNA DLC that was working fine on any of the ESX 3.0.2 hosts in this cluster. I rebuilt one to ESX 3.5 and migrated this gateway VM to it and DLC traffic stopped. I migrated it back to the original host and it works fine. I upgraded an existing ESX 3.0.2 host to 3.5 and migrated the gateway VM to it and DLC stopped. I tried opening up the ports 12000-12004 on the balky host with no success. So I stopped the firewall on the balky host, no improvement. Any thoughts as to what's different between ESX 3.0.2 and ESX 3.5 that interferes with DLC?

[divintas]

hi,

only a suggestion. Did the SNA make use of multicasting? If so check the security settings for the port group and enable promiscious mode.

Maybe the DLC 802.2 itself is the problem here. As far as I know is this protocol at the data link layer of the OSI model. Maybe someone else can post here if a non IP based protocol works with VI3.5.

regards,

[warnert]

I have the same issue, were you able to fix this?

[DHBob]

12/28/07 update. So far this seems to have stumped the band at VMware support. I tried setting the port group for promiscuous mode, no help. By the way, when I set promiscuous mode, I could not migrate my test server from a non-promiscuous host to that host. I had to migrate it without promiscuous mode set, then set it. Good to keep in mind since whenever we solve this, I'll have 11 servers to migrate.

[warnert]

Thanks for writing back, I too created a ticket with them and they were unsure as to the problem. Unfortunately I upgraded all my hosts, luckily its a development server and not production.

[DHBob]

1/8/08 update. Still nothing new from VMware tech support. We're trying to get it escalated but haven't heard anything back. Anyone else with the problem heard anything new?

[jpelland]

We do also have a bunch of guest servers with MS HIS 6 to run a legacy but very critical application. We experienced the same symptoms and went trough the same resolution steps so far without any success. I've also opened a SR. ESX 3.5 seem to have major issues with the DLC protocol while 3.0.2 was working just fine. I'll let you know when we'll have more info.

[admin]

DHBob, warnert and jpelland could you please msg me your SR numbers?

If anyone else is experiencing this issue please file a support request with us.

Thanks,

-Kris

[strandgaard]

Hi

I have the same problem. I tryede to create a SR but have a problem with that at the moment

//Allan

[DHBob]

My SR is 1100245401

[GCADAN]

We have had same situation here, error is that the remote host connection could not be reached and all PU are in pending state...would like to know if any updates so far? thanks a lot!!!

[jpelland]

Our SR is #1101068215 and the call have been escaladed to engineering last Friday. Still waiting on an update from VMWare.

[DHBob]

12/18/08 update. Engineering now has this and they have requested any SR numbers that are open and contact from anyone else that is having the problem. Our SE is working on getting a timeline for a patch.

[warnert]

Rolled back one server to 3.0.2 and DLC worked fine. My case is SR# 1100573101

[allank]

Our SR # 1101320931, we have 1 server still running 3.0.2 for det HIS server

[stuten]

Our SR# 1102088321.

We are running Windows 2003 with HIS 2000. The server has run on 3.0, 3.01 and 3.02 with no problems for months. I upgraded one node in the cluster to 3.5 and as the HIS server migrated to it, the SNA stopped working within 10 minutes. Migrate it to a pre 3.5 host and it immediately starts working, move it back to 3.5 it breaks within 10 mins, etc.

[stuten]

The SE I spoke with said that a patch is being worked on for this problem. He would not give an exact time (of course), but suggested that a patch would most likely follow with in weeks (understanding that it could still be more than 4 weeks). I was fairly encouraged by his response.

[JBanker]

I found my answer here before opening an SR - could you please notify me when the patch comes out?

Thanks,

Jim

[TarunC]

Are there any updates to this problem yet???

[jpelland]

VMWare asked us to test something to confirm the issue. Changing the vmxnet to vlance didn't work for us and we've tested with all the possible combinations. The next day, I've received the following email from Support:

Thanks for the test and response. On our end the engineers have found the reason the LLC is breaking. I will monitor the PR and advise you on when a patch is released.