VCP 5 – Objective 7.1 – Monitor ESXi, vCenter Server and Virtual Machines

Describe how Tasks and Events are viewed in vCenter Server

• Add a host
• Change the power state of a VM
• Change cluster power settings (DPM)
• Change resource settings of a resource pool or VM (CPU and Memory Shares/Reservations/Limits)
• Check the compliance of a host profile
• Create/Clone/Deploy/Export/Import a VM
• Migrate a VM (vMotion and Storage vMotion).
• Snapshot a VM
• Scan for updates and remediate an object.

There are a few rules as well on how vSphere manages tasks

• The user performing the task must have the proper permissions to do so.  If a scheduled task is created and the permissions are then removed for that user, the task will continue to run.
• When operations required by manual and scheduled task conflict the activity due first is started first.
• When a VM or host is in an incorrect state to perform the activity the task will not be performed.
• When an object is removed from the vCenter server, all associated tasks are also removed.

Events

Events are actions that occur on an object.  They include user actions and system actions.  Each event records an event message.  As with tasks you can view events on a single object or all objects in the inventory.  When you are connected directly to a host, the Tasks and Events tab is only labeled Events.  Again, as with tasks, the events shows events for the object selected as well as child objects.  Events contain a filtering option as well with the following options; Description, Type, Date, Task, Target, and User.

Identify critical performance metrics

• MCTLSZ – the amount of guest physical memory reclaimed by the balloon drivers.  A large value means that a lot of this VMs physical memory is being reclaimed to decrease the hosts memory pressure.
• SWCUR – This is the current swap usage.  Basically the current amount of physical memory being swapped out to the backing store.  This is vmkernel swapping, not guest OS swapping.  Basically means that the VMs memory is not in physical memory, but on underlying disk which is much slower.  This is not a big deal if the memory is not accessed that much, however if this value is high and SWR/s is over 0 then it is currently reading from the memory on disk.
• SWTGT – This is the expected swap usage of the VM.
• SWR/s – Number of reads from swapped memory.  This is very bad, meaning that the VM is wanting to read memory back from the swapped disk into its physical memory.  Very bad for performance.
• SWW/s – Number of writes into swapped memory. This will happen if SWTGT is greater than SWCUR.  This occurs when either the host is overcommited or the memory used by the VM reaches its memory limit of itself or its resource pool.
• ZIP/s – If this is greater than 0 then the host is currently compressing memory.  This could occur if it is overcommited.
• UNZIP/s – This indicates if the host is currently access compressed memory.  Usually indicates that the host was once overcommitted.

• %USED – Percentage of CPU time that is used by a VM.  A high value in this metric indicates that the VM is using a lot of CPU resources.
• %RDY – percentage of time that a world was ready to run, but had to wait on CPU.  Normal causes are over provisioning of CPU's to VMs.  This can also be caused if there is a CPU limit set on the VM (See MLMTD).
• %CSTP – This is the percentage of time that a VM or world spent in a ready, co-scheduled state.  This is only meaningful for VMs with multiple CPUs.  This normally means that a VM is not using multiple vCPU's in a balanced fashion.  Should either decrease CPU's or check to see if the VM is pinned to any CPUs.
• %SYS – percentage of time spent by system services on behalf of the VM.  A high value in this metric usually only indicates that the VM is very IO intensive.
• %SWPWT – Amount of time that the VM was waiting on swapped pages to be read from disk.  Usually indicates a memory problem.
• %MLMTD – Amount of time that a VM was ready to run, but was deliberately not scheduled due to violations of the CPU Limit setting.  This will also cause %RDY to increase.

• PKTTX/MBTX – Number of packets/megabits transmitted per second.
• PKTRX/MBRX – Number of packets/megabits received per second.
• DRPTX – Percentage of transmitted packets dropped.  Usually means that the network transmit performance is bad.  Could check whether physical NICs are currently using all of their capacity.  May need more physical NICs or better load balancing policies implemented.
• DRPRX – Percentage of received packets dropped.  Usually means the network is highly over utilized.  If on a VM you could try to increase the CPU resources of that VM.

• GAVG – This is the round trip latency as it appears to the VM (see KAVG and DAVG, its normally the sum of both).
• KAVG – Latency inside the vmkernel.  High KAVG often causes or is caused by queuing (see QUED)  This value should be small in comparison to DAVG, and should really be close to 0.
• DAVG – Latency as seen at the device level (HBA).  Basically the round trip time from the HBA to the storage array.  DAVG is a good indicator of a performance issue of the backend storage.
• QUED – Number of commands in the vmkernel that are currently queued.  Could be an indicator that your queue depth is set too low.  Follow your arrays instructions for queue depth settings.
• ABRTS/s – Number of commands aborted per second.  These are issued by the VM because commands are taking too long to complete.  Resets issued by the VM could also be tagged as aborts as well.  Normally caused by failed paths.

• Shows multiple charts/metrics on one page. CPU Memory, Disk and Network.
• Different metrics and charts are displayed depending on the inventory item selected.
• Allow the ability to change the view (this will differ depending on the object selected) and the date range.
• Different view options are as follows
  • Datacenter – Clusters, Storage
  • Cluster – Home, Resource Pools & VMs, Hosts
  • Resource Pool – Home, Resource Pools & VMs
  • Host – Home, VMs
  • VM – Home, Storage
• Time range can be changed to Day, Week, Month, or a custom value.
• Works well for a broad overview of an inventory object.

• Allow for more extensive analysis of an objects measurable.
• Charts can be exported/saved as jpg, bmp, gif, png, or xls.
• Can switch chart types from line and stacked graphs.
• There are really too many metrics that you can display to list here.
• Charts can be popped out into a separate window.
• Most everything is customizable and configurable, date range, intervals, metrics being measured, etc…
• Shows the latest minimum, maximum, average, and latest for the measured items.

• Receiver URL – DNS Name or IP of the receiver
• Receiver Port – Port the receiver is listening on.  If left blank it will use the default SNMP port of 162.
• Community – the community identifier.
• Optionally you can also enabled up to 3 additional receiver to a maximum of 4 total.

After this of course you would need to setup your SNMP receiver and load the VMware MIBS.

• Active Directory Timeout – timeout interval (seconds) to use when connecting to AD.
• Enable Query Limit – Limits the number of users and groups displayed in the Add Permissions box.
  • Users & Groups – Enter the number of groups/users to be displayed.  If you enter 0, all users and groups are displayed.
• Enable Validation – vCenter will periodically check it's known users & groups against active directory.
  • Validation Period – Number of minutes between synchronizations.

SMTP Settings

• SMTP Server – DNS name or IP address of the SMTP Server
• Sender Account – email address of the sender account.

• None – Turns off logging
• Error – Will only display error logging entries.
• Warnings (Errors and Warning) – Displays only warnings and errors
• Info (normal logging) – Displays information, error, and warning.
• Verbose – displays information, error, warning and verbose entries.
• Trivia – displays information, error, warning, verbose and trivia entries.

• Frequency/Start Time
  • Once – Can either then select Now or Later and enter a date/time.
  • After Startup – You then fill the delay setting in (minutes)
  • Hourly – In Start Time enter the number of minutes after the hour to start the task.  Then fill out the number of hours to run the task in the Interval .  I.E. 30/5 will run the job at half past the hour every 5 hours.
  • Daily – Enter start time and interval.  I.E.. 1:00 am/2 will run the job at 1am every 2 days.
  • Weekly – Again, Start time and interval need to be populated as well as which day(s) of the week to run.
  • Monthly – Needs Start time as well as days of the month to run by either entering the specific days of the month (dates) or by selecting the week of the month (first, second, third, forth, or last) and then selecting the day of that week.  Also, you can provide an interval to run every 1 month, every 2 months, etc.
• Email notification can be setup.

The following are the tasks that are available to be scheduled in vCenter

• VM Resources
• Host Resources
• Datastore Resources
• vMotion Resources

VM Resources allow you to map the VM to Networks and Datastores.  When selected on the cluster you can also view fault tolerance relationships.

Host Resources allow you to see the relations ships between the VMs, Networks, Datastore to that host.  

The Datastore Resources allow you to apply the same Host and VM resources but map them back to a datastore.

The vMotion maps are one of my favorite maps as they have a little bit of intelligence to them.  This map view is available when a VM is selected in the inventory and displays hosts that are compatible (green circle) and incompatible (red x) for migration targets.  It also shows the current CPU load of the host as well.