VMware vCenter Lab Manager 4 Experiencing Problems When Settings Promiscuous Mode to Enabled while using a Fenced System

I wanted to share with you a problem which affected us greatly and caused  a lot of problem. I am hoping it could help you in case you are suffering the same problem.

Problem                                                                                                                                                                                                                        We have installed a new Lab Manager 4 environment. Our environment includes several ESX servers version 4 update 1. Our working method is to work with a large number of configurations at the same time, therefore we use fencing. In addition, each configuration includes a few virtual machines with two network adapters.

We have mapped the second network adapter in each virtual machine which had two network adapters to work with Promiscuous Mode, but we have noticed that it didn’t actually worked. As part of our troubleshooting, we deployed an unfenced configuration which includes the same servers and the Promiscuous Mode worked perfectly.

 In order to solve this issue, we have taken out the second network adapter from the fencing by changing the second network adapter (in every virtual machine which has 2 network adapters) in the vCenter server into our unfenced network adapter that was defined with Promiscuous Mode. Sounds good? Not Quite… The new problem was that after that change , effective immediately another vCenter process started which undo the change that we have made, and left the second network adapter unconnected.

Solution                                                                                                                                                                       After consulting with VMware engineering support, we were asked to update a certain record in the lab manager’s database which dismisses the monitoring of the vCenter server from the lab manager’s DB.

Please be aware that this solution is highly risky; please make sure you have taken a full and updated backup of your lab manager’s database.       If you would like to cancel the monitoring of the vCenter from the lab manager server, please follow the instructions below:

1. Connect to you lab manager server.

2. Download and install: “SQL Server Management Studio Express” on the server

3. Run the following command:

UPDATE config SET VALUE=’0′ WHERE name=’NicPortgroupChangedOutsideConfigCondition’ AND CAT=’Reactions’

Idit.

Read Full Post »

ESX 3.5 update 5 – Server Stuck During Startup On “vmkernal lpfc_740” Module

We have received a new server which we installed on it ESX 3.5 update 5. After configuring the external storage for the server (we uses SAN storage array) during startup, the server was stuck on “vmkernal lpfc_740” Module. We didn’t understood what might be the problem.

We tried to start the server each time with a different HBA card and one HBA card worked fine and the other was stuck on the same module. We even thought that the OS is corrupted and reinstalled the server, but the problem still existed.

We thought that the slot might be damaged and we changed the HBA cards location, the good HBA card still worked fine and the other not. So we assumed that the HBA driver wasn’t up to date, but after a quick check we saw that the driver was up to date as well.

Solution                                                                                                                                                                                                           Afterwards we saw that the server’s problematic HBA card was defined on a different SP in its zoning. Each time we removed a different SP from its zoning, and found out that we have a problem with only one SP. After changing the FC cable from the FC switch into the SP port in our storage array , it worked great!!

What’s amazing is that even though the SP port in the FC switch had a link and the SP switch saw the port correctly, the problem was with the FC cable. (and it was fresh from the bag). After changing the cable, the problem solved and the server was started quickly. And of cours the server saw its storage.

Conclusion – Always be suspicious regarding new purchasing.

Idit 🙂

Read Full Post »