Previously I’ve got no chance to configure a NIC teaming on my ESX host because to do this I need at least 2x network cards before I can do load balance for each VMs traffic. Luckily, while I’m cleaning up my store room yesterday, I found old 3com 3c905c-TX NIC which off-course supported in VMware HCL. After fixed it into my Asus M2A-VM motherboard and ESX was able to detected it automatically without any modification on 3com pciid.

So what is NIC Teaming actually?

NIC teaming is the process of grouping together several physical NICs into one single logical NIC, which can be used for network fault tolerance and transmit load balance. The process of grouping NICs is called teaming. Teaming has two purposes:

• Fault Tolerance: By teaming more than one physical NIC to a logical NIC, high availability is maximized. Even if one NIC fails, the network connection does not cease and continues to operate on other NICs. Read more

I just spent 6 hours today to troubleshoot one of my customer SLES 10 virtual machine which able to boot the kernel successfully but forever stuck on the stage “waiting mandatory devices as per attachment below. The ESX host previously has been powered off due to power failure and this caused SLES 10 network config for eth1 & eth2 devices id changed from :

eth-id-:00:50:56:b5:78:d8 to

eth-bus-pci-0000:00:11.0

Somehow I managed to troubleshoot this issue with below method :

• Boot the vm with Sles 10 or any live-cd .iso
• Enter “rescue mode”
• Login as root
• Mount the LVM volume & cd to /etc/sysconfig/network
• edit config file and look for MANDATORY_DEVICES='’to become (MANDATORY_DEVICES=’eth-id-00:50:56:b5:78:d8′)
• Reboot SLES 10

Now SLES 10 vm able to boot successfully without problem.

ariyossss

athlon_crazy 27032009 11:09PM

Since I already published this tools for download in VMware Blogs, I will do the same thing here.

If you still don’t have any backup tools for your home or medium scale production ESX to backup your virtual machines, perhaps you can try this small tools to do backup task for you.

REQUIREMENT
vcb-1.0.5
vcbMC-1.0.6 Beta Download –> (Asian), (Euro) - Credit to Jean Charles..
Windows Server 2003
CMD.exe
Windows Task Scheduler Read more

Since I lost my Scalix virtual machine last time, I’ve started to have my weekly virtual machine backup with vcb. I admit that vcb can do a pretty good job to backup all my virtual machines when I’m using it with vcbMC, but still it’s not good enough for me simply because :

• It’s running on windows platform
• Time consuming during restoration
• Proof of concept

Since out of 14 vm’s that I have right now running on Linux platform, to eliminate my single windows VM is my main objective. Restoration of vm require me to upload the backup image to the NFS datastore first before I can start to restore it. So I’m start looking for another option which is fast, less downtime, running on Linux platform and that’s how DRBD came into my mind.

So what DRBD can do and how DRBD can help me? :

• Linux platform
• Mirroring NFS (Datastore)
• Running very well with No-x Linux
• Linux-HA integration

DRBD doesn’t need an identical machine except for the storage to do mirroring and it can be run on any Linux distro but of-course, I have to use it with proper Linux Kernel. Since I already managed to recompile No-x Linux kernel with DRBD version 8.3.0, my job is become easier.

Before I buy another physical machine to do mirroring for NFS datastore, I will do some testing on 2x vm’s with below setup : Read more

We are currently conducting 2 days JumpStart training for our customers which held in Cyberjaya. Basically this training will cover :-

• Virtualization Introduction
• ESX installation (ESX Installation)
• Virtual Center (Management)
• VMware HA, DRS & VMotion

My colleague KC doing briefing to the customers..

ariyossss

athlon_crazy 23/03/2009 10:56 am

After few research & three(3) times I recompiled ESX/redhat kernel, finally I managed to fully utilized my IDE hdd inside my ESX server this morning. As we already knew by default :

• ESX doesn’t allow local IDE to be a Datastore
• ESX come with NFS server version 3 on udp
• ESX only support NFS mount for NFS server version 3 on tcp

My setup currently :

• Asus M2A-VM without supported SATA controller
• VMware ESX 3.5 U3 installed on IDE
• Datastore using NFS on another physical Linux machine

My Objectives :

• Eliminate physical Linux machine (Perhaps will sell it later)
• To have local Datastore

How To :

I just followed a guideline as per given url (Thanks a lot Derek)

Here

Notes :

Follow the guideline very carefully and don’t miss any step. (NFS Only)

Result :

ariyossss

athlon_crazy 22/03/2009 9:56 AM

Just wan’t to share with those who having Endian Firewall in their current environment and their HTTPS is not working on Transparent Proxy setup. There’s nothing to do with your transparent proxy setup since we really can’t redirect https encrypted packets to another port.
Scenario :

• Endian Proxy + authenticate + Browser setting https port 8080 = http & https (OK)
• Endian Proxy + transparent = http (OK) & https (KO)

For those having this issue you probably can do as below :

• Endian 2.2 rc1 = $iptables -I FORWARD -j ACCEPT
• Endian 2.2 rc3 = $iptables -t nat -I POSTROUTING -o br0 -j MASQURADE

Both method has been used and tested working on my & customers Endian Firewall.

Thanks

ariyossss….

athlon_crazy 21:02PM