You may have read or heard that VMware will disable the ability to share memory pages “between” virtual machines (Inter-VM Transparent Page Sharing) by default (in ESXi 5.0/5.1 and 5.5) in coming updates and the next major ESXi release.
In some environments this may lead to an increased host memory demand to support the same number of workloads.
VMware describes the reasons for this decision in the VMware KB 2080735: Security considerations and disallowing inter-Virtual Machine Transparent Page Sharing
In order zu find out if this change impacts your environment you can use the “Host Memory Assessment Tool” scripted by Brian Graf (@vBrianGraf) and Mark Achtemichuk (@vmMarkA).
You can download and read more about the tool in the VMware vSphere Blog: Assess the Performance Impact of the Security Change in Transparent Page Sharing Behaviour
To run the tool you need credentials to access the vCenter Server and root access to the ESXi hosts.
Other requirements are:
- PowerShell 3
- PowerCLI 5.5 or higher
- plink.exe in c:\temp (you can download plink.exe here)
- read/write access to C:\temp for working files
How to use the tool:
After entering the credentials for vCenter and the ESXi hosts click “Connect”.
The tool will now list all hosts in your vCenter and the SSH Status (running or stopped)
If necessary you can now select dedicated hosts and enable or disable SSH.
To run the assessment for dedicated hosts select the hosts (check-box) and click “Select”
Depending on the number of hosts it will take some time – then you will get the following information:
- Total Host memory (GB):
amount of physical memory available for the VMKernel
- Host Mem Saved via TPS (GB)
amount of memory TPS has currently saved
- Host Mem Saved via TPS Zero Pages (GB)
amount of TPS saved memory that are zero pages
- Potential Host Mem Savings Lost (GB)
difference of total TPS saved memory minus the zero pages. This is the amount of memory you will need more when Inter-VM TPS is disabled
- Host Free Mem
amount of memory currently being reported as free by the VMKernel
If “Potential Host Mem Saving Lost” is larger than the “Host Free Mem” value you might get a problem (eg. swapping, ballooning).
It took a long time, but with patch ESXi-5.5.0-20150104001-standard (2099271) VMware managed to fix the Changed Block Tracking (CBT) Bug for ESXi 5.5 together with a number of other known issues.
Here is an excerpt from the release note:
When you use backup software that uses the Virtual Disk Development Kit (VDDK) API call QueryChangedDiskAreas(), the list of allocated disk sectors returned might be incorrect and incremental backups might appear to be corrupt or missing. A message similar to the following is written to vmware.log:
DISKLIB-CTK: Resized change tracking block size from XXX to YYY
For more information, see KB 2090639.
A fix for ESXi 5.0 and 5.1 is available since December, 15th: Changed Block Tracking (CBT) Bug fixed for ESXi 5.0 and ESXi 5.1
Up to version Veeam ONE 7 the account used for connecting Veeam ONE to a vCenter only required „Read-only“ permissions as a minimum.
This has changed with Veeam ONE v8.
For example, the software now uses the new CIM-SMASH protocol to collect hardware sensors data.
The account used for connection to vCenter must have at least the permissions as described in the How-to below.
How to edit a role with minimum rights for Veeam ONE v8:
- open the vSphere Web Client
- change to Home – Administration – Roles
- right-click the Read-only role and choose „Clone…“
- edit the role name (eg. Veeam_v8)
Now add the following permissions to the new role:
- Host – CIM – CIM Interaction
- Host – Configuration – Connection
- VirtualMachine – Interaction – AnswerQuestion
- VirtualMachine – Interaction – Console Interaction
- Global – Manage custom attributes
- Global – Set custom attributes
- Global – Licenses
- Datastore – Browse datastore
Save the settings and apply the new role to the vCenter User – done!
I just upgraded Veeam ONE from 7 to v8. The upgrade is as easy as it is known from Veeam.
There is only one scenario you should take care of: if your Veeam ONE 7 SQL database has grown over 10 GB.
During installation a procedure is triggered that rebuilds indexes in the database. They do this to boost the report generation performance and other database queries.
But if your database is larger than 10 GB, this procedure needs a lot of time to complete. In this case you will see this window for a longer time:
Installing Veeam ONE Monitor Server…
Veeam has published a Knowledge Base article (KB 1951) describing what you can do if your Veeam ONE SQL database is larger than 10 GB:
Upgrading Veeam ONE to version 8 in larger environments
If you start the upgrade without reading/taking into account the KB article, do not panic.
The time required for the script to complete depends on the database size.
All you need is time – when the rebuild of indexes is finished, you can proceed with the upgrade to Veeam ONE v8.