Hi there! Today I would like to talk about differences between VmWare`s hypervizor and OS level clusterization. So, let`s begin: Read the rest of this entry »
In Scott Herold`s blog I found a very interesting article: How to remove FT from host without vCenter? Yes, I know, that it`s almost an unbelievable situation, but…
Genius of Mark Russinovich has released a new tool – PC2VHD. This tool can convert running Windows to VHD file (I bet that it will never create VMDK 
). It uses VSS in Windows for VHD files creation. At the same time, Paragon Software has released its software for P2V and V2V convertion – Virtualization Manager 2009 Corporate. This tool is more powerful than disc2VHD (but Mark`s tools are always simple and small). I`ll test it soon, and will post my results here.
Let’s have a quick look at password policy in ESX Server:
esxcfg-auth –usepamqc=N1 N2 N3 N4 N5 match
where N1 is minimal length of password, which consists only of 1 class symbols (abcdef….)
where N2 is minimal length of password, which consists of 1 class symbols or 2 class symbols (ABCdef…)
where N3 is minimal amount of letters in password
where N4 is minimal length of password, which consists of 1, 2 and 3 class symbols (ABCdef123)
where N5 is minimal length of password, which consists of 4 class symbols (first 3 classes and tech symbols !@#$%^&*)
where MATCH - is a number, where you can`t repeat symbols. For example, parameter 3 will make password 112233 unusable.
If any option will have parameter -1 it will be turned off.
Maximum amount of attempts for logging in can be set with function esxcfg-auth –maxfailedlogins=. After exceeding this number the account will be blocked. You can see amount of failed authentications by faillog -u user, and reset it by faillog -r vpxuser.
Everything is really simple! I need only 30 seconds to show it! 
esxcfg-auth [ [ --enablead | --disablead ] [ --addomain= | --removedomain= ] [--addc= | --removedc= ] ]
where addomain|removedomain is name of domain, and addc|removedc is DNS or IP address of domain controller. After that you need to create a user, who will be able to connect to service console, using useradd command. That`s it!
Today, in the 3d party of my “In a nutshell” story I`ll talk about VmWare High Availability: what it is, and how does it work.
Hi there! Today I`ll continue talking about disaster recovery functions, provided by VmWare, and I`d like to cover question of requirements and limitations of Fault Tolerance technology. As always, I`ll be terse and clear. Read the rest of this entry »
What else can I say?! Lots of new features, at last Windows 7 is supported, and Player can create VMs now! Download here – http://communities.vmware.com/community/beta/workstation
Hi there! Today I`d like to offer you an article about high availability technologies, provided by VmWare: vMotion, HA and FT. vMotion - the technology allows to transfer VMs from one physical server to another while it`s working (analog of Live Migration in Windows Server Hyper-V R2). Also need to mention that vMotion uses advantages of DRS (Dynamic Resource Scheduler), which allows load balancing.
Storage vMotion - allows transferring VMDK file of VM to another storage device while VM is running.
Difference between these technologies is that first one allows turning off all servers, if VMDK is on shared storage, while the second one – moving VMDK file with VM to another storage, and turning it off.
VmWare HA - this technology scans hosts and VMs for availability all the time. If hosts go down, then VM manager runs them on another physical server. It`s like a hard reboot for VM.
Fault Tolerance - disaster recovery function, which allows creating 24/7/365 running virtual machines. Works as a usual cluster: two identical VMs are running on two physical servers: one node is running in active mode, while second – in passive. All processor`s commands are one side replicated between VMs. Difference between FT and HA is the following: as I said, running VM in first case on another host is equivalent to hard reboot, while in case of FT second node is running from the moment the first node goes down. Need to note that using FT has some limitations: first, both hosts MUST have identical CPUs and second – list of supported CPUs is limited.
That`s all for today. In the next article I`ll tell you how to install and configure all these features. See you!
Today I would like to talk about difference in various types of vNIC in ESX 4. Need to note that this device have no physical analogues, and can be used only if VmWare Tools are installed. VMXNET - first generation, optimized for network performance between VMs.
VMXNET2 or Enhanced - simple VMXNET, but with few additional features – Jumbo Frames (only in 10Gb networks) and TCP Segmentation Offloading (TSO). This vNIC is available beginning from ESX 3. It is also worth mentioning that you need to read this article in VmWare`s Knowledge Base for using this adapter in Windows 2003 (not Enterpise and datacenter editions).
VMXNET3 - next generation of vNIC, now it uses paravirtualization (in a nutshell – earlier everything was virtualized, now parts of commands are sent directly to pNIC). VMXNET3 is not a subset of vNICs of previous generations, but supports all of their functions. It is available only in VMs with virtual hardware v7 and higher.
Now, when we`ve classified types of vNICs, let`s see what kinds of adapters are used in VMs:
Vlance - virtual analog of AMD 79C970 PCnet32 LANCE NIC. Drivers are available in all OSes till Vista. Used in VMs without VmTools.
Flexible - while OS is booting, it operates as Vlance. If VmTools are installed on VM, it will operate as VMXNET, if not – as Vlance.
E1000 – is an emulated analog of Intel 82545EM Gigabit Ethernet NIC.
In addition, I would like to list all functions available in all vNICs in the table below:
|
Flexible |
Enchanced vmxnet |
E1000 |
vmxnet3 |
|
| IPv4 TSO |
- |
+ |
+ |
+ |
| IPv6 TSO |
- |
- |
- |
+ |
| Jumbo Frames |
- |
+ |
- |
+ |
| Large Ring Sizes |
- |
- |
+ |
+ |
| RSS |
- |
- |
- |
+ |
| MSI-X |
- |
- |
- |
+ |
| Virtual hardware version |
4 or 7 |
4 or 7 |
4 or 7 |
Only7 |
Hi there! Just for beginning pros and cons of using VmWare Server instead of ESXi Server:
Pros:
- Server doesn`t require compatible hardware, and can be installed on any Windows or Linux host
- ESXi doesn`t allow pass-through of USB drives to VM
- Server`s VMs can be copied easily to USB stick and moved to other host
- Doesn`t require additional learning.
Cons:
- Requires too many RAM memories for Apache web server – about 240 MB.
The easiest way to avoid this is to use VMware vSphere Client, which can be obtained with VMware vCenter. All you have to do after vSphere Client installation – put string type server_address:8883 in IP address, and you`ll be able to use all advantages of both products – vSphere Client and VmWare Server.
Hi there! Today I would like to tell you how to simply backup your VMs to a remote tape device. Let`s say we have the following configuration: 192.168.1.100 – Windows Server with tape device, 192.168.1.200 – ESX host, 192.168.1.201 and .202 – Windows based VMs. In this configuration we do not have to work with ESX at all. Additionally such separation of IP zones has been made especially to simplify reading. Read the rest of this entry »
During last few months I`ve noticed that people don`t see any difference between Rocket Division and StarWind Software company. Yes, you got all right – formerly StarWind software was developed by Rocket Division, but beginning from February 2009 StarWind iSCSI SAN software has been developed by StarWind Software. The idea is simple – focusing on developing of such emerging and potential technology as iSCSI. Need to mention that Rocket Division continues to develop its burning software StarBurn, and StarBurn SDK.
So, the conclusion is: Rocket Division continues development of burning software and SDK called StarBurn and StarBurn SDK respectively, while new company – StarWind Software, develops iSCSI SAN virtualization software and initiator, called StarWind and StarPort respectively. Moreover, we can already see benefits of such an approach: as I was told, next major version of StarWind will support High Availability configurations, while StarPort will be not only an iSCSI, but also FCoE (Fiber Channel over Ethernet) initiator (as far as I know it will be the first software FCoE initiator, because DataCore had only announced it for future versions of its software).
P.S. Thanks a lot to Rocket Division and StarWind Software companies for help and information!
P.P.S. I forgot to say that StarWind Software also develops V2V Converter – a tool for VHD to VMDK and vice versa images converting.
Hi there. Today on VmWare blogs I found that VmWare buys out SpringSource company, that is the biggest contributor and maintainer of Apache Tomcat and developed Spring Framework, together with a lot of other Java-based offers. The main goal of this is to lead virtualization and cloud computing to a new level, and develop fully-integrated PaaS (Platform as a Service).
Hi there! Sorry, got no time to write. Few days ago I got an idea about series of Versus articles. The first one is VmWare ESXi vs MS Hyper-V, and the second one will be iSCSI: hardware vs software.
All my thoughts about VmWare vs MS are available under cut. So,VmWare`s footprint – 100 Mb, while Hyper-V is a version of MS Server Core, that`s why it requires more than 1.5 Gb.
ESXi is an independent solution, and that means no 3d party drivers and good stability with less amount of patches. Hyper-V is a part of WS2k8 Core, with a lot of 3d party drivers and a lot of monthly updates that require reboot.
VMotion vs QuickMotion – they are competitors.
VMFS is a clustered file-system, while NTFS is not. The only way out – 3d party iSCSI target (Who mentioned Starwind iSCSI?)
VmWare allows memory overcommitment, while MS doesn`t even think about implementing it. At least now, and possibly, will not ever.
Oh, and I also have to note that I`m talking about VmWare ESXi 3.5 and MS Hyper-V (no R2).