Previously we saw tutorial about Dynamic Memory in Hyper-V. Today, we will see Smart Paging in Hyper-V. We strongly advise you to consult this tutorial before coming to it in order to better understand it.
We will now see what exactly Smart Paging is and its contribution in this explanatory tutorial.
For information Smart Paging is also called as “intelligent pagination”. It is a mechanism to counteract the difficulties that may be encountered during the restart of a machine, if the Hyper-V is not able to assign VM to its minimum RAM which required for starting a machine (Startup RAM).
This is due to the fact that due to the dynamic memory allocation, VMs have their memory threshold which varies from each other. Therefore, when you try to restart a VM and if there is not enough memory available and the Hyper-V host is also not able to recover minimum required RAM from the other VM, the VM cannot start. Fortunately, now we have Smart Paging feature.
Indeed, in this case a Smart paging file will be created and used to virtual machine time that Hyper-V host can allocate the minimum RAM which is necessary. Once the RAM allocation carried out, the Smart Paging file will be automatically deleted. Moreover, it will be stored at the location indicated in the VM’s properties.
To unload the VM memory allocated through Smart Paging, technology “Memory Ballooning” is used as in the case of DRAM via Windows Memory Manager. Note that Smart Paging is a temporary mechanism. This allows the virtual machine to start but this memory cannot be used indefinitely to the virtual machine.
Besides, Microsoft recommends not using this feature more than 10 minutes on a machine while specifying RAM – as we said – that this in only temporary.
The Smart Paging can be used for restarting a VM or while resetting the VM. And if there are no VMs or physical RAM available, then it is not possible to recover the RAM on the other VMs.
Hyper-V version 3 with Windows Server 2012 has a new feature which dynamically allocates memory (RAM) on a Windows virtual machine. What is that? And is it a complicated process?
It is simple; it adds memory to a VM as needed up to a limit defined in the configuration. The dynamic addition is actually a hot-add memory to the VM.
With this feature we can more efficiently manage the memory allocation to the virtual machine since without this feature we must first turn off the VM to change the allocated memory. With this feature, everything is handled dynamically. That’s to say hot-add by the Hyper-V manager, which interact with the virtual hosts.
Hyper-V also incorporated this feature in Windows Server 2008 R2 with Service Pack1. The following guest operating systems are supported:
We will see a brief example here about adding and deleting memory dynamically on a VM to understand how it works…
Our Hyper-V server contains a VM that uses dynamic memory feature. It has 2 GB of memory assigned minimum and begins to saturate.
After enabling the dynamic memory feature, Windows will show warning message that the hypervisor needs additional memory through dynamic memory driver: DMVSC, Dynamic Memory VSC.
You will clearly understand in order to use this feature requires the system that taking advantage of the compatible virtual host which supports dynamic memory driver.
OK, now our VM has a 3 GB of RAM, thanks to hot-added memory. The hours pass…the peak load of activity decrease and then our VM does not need as much memory.
A signal is sent to the Hyper-V host to withdraw the memory from our virtual machine, but we cannot remove memory directly from a Windows machine enthusiastically, otherwise a crash will obtain with a famous blue screen.
To counter this problem there is a solution. In fact, Hyper-V does believe in the system of exploitation when the VM memory is unavailable. Alright, as our VM no longer need to use that part of memory, it is then possible to reassign it to another needy VM. This is called as the “Memory Ballooning”.
The added extra memory to a virtual machine, 1GB, will be rendered in the virtual machine. We will allow it to use only minimum 2 GB.
To enable dynamic memory feature on virtual memory in Hyper-V manager, right click on the VM on which you need additional RAM, then click on “Settings”. Then on the left side, click “Memory” and in the right pane, select the check box next to “Enable Dynamic Memory”.
However, there are various fields:
-Minimum RAM: it must always be made available to the VM, and then the VM memory will not drop below this value. This is the minimum memory that the VM needs to function.
-Buffer: This buffer will be allocated to the VM memory very quickly with an area pre-allocated memory on each VM that can react quickly during a peak load.
It may be interesting to define a high value for this field when the VM is running an application that require a large amount RAM chunk for a moment and then release afterwards. This will meet the needs of the application. Conversely, set a low value on a VM if it turns an application that always seeks to use the maximum possible memory.
To determine this value, Hyper-V uses the percentage you point it and the amount of memory used at a time T by the virtual machine. After a calculation, it shows a value that will correspond to the available buffer.
-Memory Weight: It manages the memory availability on the priority of a machine relative to another. The higher the priority, the more Hyper-V will work to prioritize allocating memory for that VM rather than another whose priority is lower. It is important to prioritize critical virtual machines first and have significant advantage in term of memory.
Cpanel server keeps frequently visited files and web traffic in log files, in these servers, locations of common log files is almost fixed for almost all versions of cPanel, hence it gets easy to track it for any service running on cpanel based server. Checkout some logs for varied services running on cPanel based servers.
Common cPanel logs
Introduction: Contains logs related to cPanel, all web traffic is logged to cPanel and webmail. Installation log contains records and errors during cPanel installation. Access log contains traffic details to cPanel and webmail. Error Log comprises of all errors that occurred while accessing cPanel. Bandwidth log contains Bandwidth history with user details.
Introduction: It is a web server managed by cPanel typically for Linux servers, it carries log for all sites accessed with cpanel. Access log contains all the details of the requests to the servers with IPs, usernames, etc. as required. All Apache errors are logged in Error log. Domain access and error logs for each registered domain with cPanel are logged in domlog.
Introduction: FTP does not have its own log file; it is connected with Message log of system and hence all details of FTP processes like login/access or failure, transactions or communications are logged in messages of the system.
Introduction: It contains general information and errors related with MySQL access with the cPanel. The address may vary but many of the times remain same.
Introduction: It is a Mail Transfer Agent used by cPanel. Exim has importantly 2 useful logs they are Main log and Reject log. Main log contains all transaction details of mails with incoming and outgoing specifics. Reject log contains all failed transaction details. And IMAP logs have all general details of transaction, failures, etc.
These are some of the important log details; however there are many other log details also. With use of these paths, we can easily access common log files in cPanel for any services running on cPanel based servers. Share other log files you know.
We are glad to announce the release of VMM-2.1.1 eNlight update which has some bug fixes and stability improvements, better user experience and faster VM provisioning. If you have any questions, comments or suggestion about your experience then please contact technical support right away.
We look forward to improve your experience on eNlight cloud hosting platform and wish to provide you enhancement with our continuous research and development.
VMM-2.1.1 – eNlight update [ Release Highlights ]
The All New iNtelligent Billing Predictor : Forecast resources usage for both the Static and Dynamic VMs.
Dynamic VM (DVM)
Static VM (SVM)
RESOURCES FOR CPU AND MEMORY:
In DVM, resource allocation is on demand and charges applied are for what is consumed i.e. pay per consume. Computing resources like Memory and CPU can be set with minimum and maximum thresholds, as shown in image. Scaling of resources is done between these limits as per requirement in virtual machine. Take an example; if the load has peaked to the upper threshold, such that it requires resources to be scaled to the maximum limit and with decrease in load resources would be scaled down automatically.
In SVM, Resource allocation is fixed and charges applied are also fixed for allocated resources as shown in image. System cannot scale up compute resources automatically.
From above descriptions, we can conclude that DVM is better compared to SVM in resource utilization; it is also cost effective solution.
In DVM, automatic vertical scaling is possible. Resources can scale up during peak hours till maximum limit provided and then scale down in off-peak hours till the minimum limit. Resources can be changed between the limits as required with increase and decrease of capacities is called as vertical scaling.
In SVM, no vertical auto scaling feature is possible, as allocated resources are fixed. If the allocated resourced needs to be edited, system needs to be shut down and start again with new allocated resources.
DVM is very beneficial in Peak load hours and due to auto-scalability feature system works in continuity unlike SVM. Thus provides, flexibility, continuity and productivity.
PAY PER CONSUME:
In DVM, charges are applied as per consumed bandwidth, memory and CPU.
In SVM, charges are applied as per consumed bandwidth, but fixed charges for allocated resources.
During off-peak hours charges applied are cost efficient in DVM compared to SVM.
PAY PER USE:
In DVM, you will be charged only for resources you have used.
In SVM, you will be charged complete for fixed allocated resources and as per utilized bandwidth.
DVM is cost effective as compared to SVM with resourceful ability of auto-scaling.
From all the conclusions drawn above, it clearly points that DVM are beneficial over SVM.