Configuration of VmgOSMemoryTest

Memory shortage on a VM can affect the memory allocation to crucial processes that are being executed on the VM, which in turn can adversely impact the performance of the applications running on the VM. One of the primary reasons for Memory shortage is that of precious memory space being unnecessarily hogged by Modified and Standby memory lists that hold temporary/unused data. The Modified and Standby memory lists cache temporary data when the applications/services run in the VM. These temporary data will no longer be used by the applications/services thus hogging memory space unnecessarily. Therefore, administrators should clear the cached data on a regular basis. If the cached data is not cleared regularly, sometimes, you may not be able to allocate memory to the business-critical processes, which will seriously impact service/application delivery and impair user experience. Therefore, it is imperative that you should closely observe if the memory shortage occurs due to data growth of the Modified and Standby memory lists and proactively initiate remedial actions before it causes severe memory contention on the VM. This can be achieved with the help of the VmgOSMemoryTest.

This test auto discovers the VMs on the virtual server and periodically monitors the memory usage of each VM, checks whether adequate physical memory is available to the VM, and if not, promptly alerts users to the same. In the process, the test also reveals the VMs on which the memory space is abnormally hogged by Modified or Standby memory list. This way, VMs that experience potential memory contention are brought to your attention. Besides warning you of memory contention that Modified/Standby memory lists can cause, the test also empowers you to avoid probable memory shortage by initiating automated actions. These automated actions can be closely tracked using detailed diagnostics.

The default parameters associated with this test are:

• The TEST PERIOD list box helps the user to decide how often this test needs to be executed. 

• In the HOST text box, the host name of the server for which the test is to be configured has to be specified.

• Specify the port at which the specified HOST listens in the PORT text box . By default, this is NULL.

• Specify the domain within which the virtual guests reside in the DOMAIN text box. Since the Solaris LDoms server supports only Solaris and Linux guests, this parameter should always be set to none.

• This test connects to each virtual guest and collects status and resource usage statistics from the guest. In order to do so, the test must be configured with user privileges that allow a remote connection to the virtual guest from the Solaris host. If a single user has access to all the guest domains on the Solaris server, specify the name of that user against ADMIN USER, and specify his password against ADMIN PASSWORD. On the other hand, if the user credentials vary from one guest to another, then multiple ADMIN USERs and ADMIN PASSWORDs might have to be specified for every Solaris LDoms server being monitored. To help administrators provide these user details quickly and easily, the eG administrative interface embeds a special configuration page. To access this page, simply click on the Click here hyperlink that appears just above the parameters of this test in the test configuration page.

• The password of the ADMIN USER needs to be provided in the ADMIN PASSWORD.

• Confirm the password by retyping it in the CONFIRM PASSWORD tex box.

• Typically, when performing VM discovery, the eG agent automatically discovers the operating system on which every VM runs, and all the IP addresses that each VM supports. If two are more VMs on a target vSphere server are in a VM cluster, then the eG agent will also auto-discover the cluster IP address. Since the cluster IP address is shared by all VMs in the cluster, this IP address will be in the discovery list of every VM in the cluster. In this case, if the eG agent attempts to obtain the 'inside view' of each VM in a cluster using their cluster IP address, incorrect metrics may be reported sometimes. To avoid this, you may want to instruct the eG agent to not use the cluster IP address when collecting 'inside view' metrics. For this, specify a comma-separated list of cluster IP addresses to be excluded in the EXCLUDE IP text box.

• By default, the detailed diagnosis of the Used physical memory measure of this test reports the number of instances of each process running on the VM, and the aggregated memory usage (in MB and %) of every process across all its instances. For example, if users to a VM are together having 15 instances of Chrome open on the machine at around the same time, then the detailed diagnosis of the Used physical memory measure will compute and display the collective memory usage of all 15 instances against the Application Name, Chrome. From this, you can quickly identify the exact process that is ‘collectively’ (i.e., across its instances) over-utilizing the memory. Sometimes, administrators might want to isolate not just the process, but also similar process arguments that are guilty of abnormal memory consumption. This granular insight will take administrators closer to the root-cause of the memory bottleneck on a desktop. For instance, in the Chrome example above, say 8 of the 15 instances are used to access the same YouTube video, and 7 instances are accessing a shopping site. In such a case, if memory usage is aggregated at the URL-level and not the process-level, then administrators can quickly identify which precise URL is draining memory - the YouTube video? or the shopping site? For this, detailed diagnostics should be grouped by process arguments (eg., URLs) and not just by process/application names. To enable grouping by arguments, set the GROUP PROCESSES WITH ARGUMENTS flag to Yes. By default, this flag is set to No.

• By default, MEMORY UTILIZATION IN PERCENT parameter is set to 75 percentage indicating that by default this test will display top-10 memory intensive processes on a VM in the detailed diagnosis of the Physical memory utilized measure only when the memory utilization reaches/exceeds 75 percentage. On the other hand, if the memory utilization is lower than the value specified against this parameter, then, this test displays only topv3 memory intensive processes in the detailed diagnosis of the Physical memory utilized measure so as to conserve storage space on the database. However, this setting can be overridden if required.

• A memory shortage caused due to memory space hogged by modified and standby memory lists has become an increasingly frequent and critical problem. The memory shortage can make a server stutter or run more slowly than usual. Such operational interruptions can make you spend more time on waiting for pages and programs to load, disturbing your work flow and leading to work delays. Moreover, memory shortage can not only impair user productivity, but can also bring crucial business processes to a halt, and can severely degrade overall user experience with the server. To avoid such an outcome, administrators need to rapidly empty the memory space unnecessarily hogged by the modified and standby memory lists. This way, administrators can ensure that adequate memory is always available for executing business-critical processes.

  To enable administrators to achieve this, eG Enterprise offers ‘Automation Actions’. These actions are governed by the ‘Automation Action Enabled’ flag.

  AUTOMATED ACTION ENABLED flag, if enabled, automatically initiates user-configured actions, so that the memory hogged by modified and standby memory lists can be cleared when memory utilization crosses a configured limit.

  Note:

  The ‘Automation Action Enabled’ parameters, once configured, will apply only to the chosen component. These settings will not be applied to other components even if the Apply to other components button is clicked in the Specific Test Configuration page.

• EMPTY MODIFIED MEMORY flag appears only if the Automated Action Enabled flag is enabled.

  Modified memory contains cached data and code that is not actively in use by processes, the system and the system cache. This memory needs to be written out before it will be available for allocation to a process or for system use. You can easily clear this memory using the EMPTY MODIFIED MEMORY flag.

  By default, EMPTY MODIFIED MEMORY flag is set to No. When this flag is set Yes, amount of memory assigned to the modified list will be automatically cleared if the memory utilization reaches/exceeds the value specified against the Automation Memory Limit parameter.

• EMPTY STANDBY MEMORY flag appears only if the Automated Action Enabled flag is enabled.

  Standby memory is a page of physical memory that is not actively used. When writing a series of frequently used information in the memory area, the programs and services running on the server will use part of memory as cache. The cache should be released and made available after executing programs and services. If the standby memory is not cleared on a regular basis, it can slow down the server operations. Therefore, it is necessary to clear the standby memory at regular intervals. This can easily done using the EMPTY STANDBY MEMORY flag.

  By default, EMPTY STANDBY MEMORY flag is set to No. When this flag is set Yes, memory that is allocated to the standby page list will be automatically cleared if the memory utilization reaches/exceeds the value specified against the Automation Memory Limit parameter.

• AUTOMATION MEMORY LIMIT flag appears only if the Automated Action Enabled flag is enabled.

  In AUTOMATION MEMORY LIMIT text box, specify the upper limit for memory utilization in percentage beyond which this test should trigger user-configured automation actions. By default, 95 percentage is set against this parameter. When the memory utilization reaches/exceeds the value specified in this text box, the user-configured actions will be automatically triggered. However, you can override this settings if required.

• The DD FREQUENCY refers to the frequency with which detailed diagnosis measures are to be generated for this test. The default is 1:1. This indicates that, by default, detailed measures will be generated every time this test runs, and also every time the test detects a problem. You can modify this frequency, if you so desire. Also, if you intend to disable the detailed diagnosis capability for this test, you can do so by specifying none against DD FREQUENCY.

• To make diagnosis more efficient and accurate, the eG system embeds an optional detailed diagnostic capability. With this capability, the eG agents can be configured to run detailed, more elaborate tests as and when specific problems are detected. To enable the detailed diagnosis capability of this test, by default, for a particular server, choose the On option against DETAILED DIAGNOSIS. To disable the capability, click on the Off option.

  The option to selectively enable/disable the detailed diagnosis capability will be available only if the following conditions are fulfilled:

  • The eG manager license should allow the detailed diagnosis capability

  • Both the bad and normal frequencies configured for the detailed diagnosis measures should not be 0.

When changing the configuration for specific servers, a “*” beside the text box corresponding to the parameter signifies that these values have to be manually configured by the user. The parameter values that require to be configured will typically be prefixed with a “$” or contain a series of “*”. A value of "none" in the parameter value indicates that the corresponding parameter value can be changed if required.