Yesterday I observed a small problem in my environment. When I clicked on the hardware status tab in the C# client I got the following error message.

I done a little research on Google and found some VMware KB articles with similar problems. In all of these articles the CIM service on the ESXi host should be restarted. I checked if the Health status on a single ESXi host is also not working but I got this:

So Health status is working on the host. Another problem all of this articles stated was that maybe the ports for the CIM service are not open or blocked be a firewall. I also checked this. ESXi was configured to allow CIM ports and the VCSA also didn’t block any ports. Both, the ESXi and VCSA, are running on the same subnet so there is also no firewall between them.

But what can cause this problem. After a short tweet, Frank Büchsel (@fbuechsel) gave me some hint where I can look for errors. When using VCSA you can find a log called vws.log (Common Interface Model log) under /var/log/vmware-vpx/vws.log. There I found the following error:

Caused by: sun.security.validator.ValidatorException: PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: unable to find valid certification path to requested target

It looks like I had an certificate problem, but why? Seconds later I had my answer. I have done a VCSA upgrade from 5.1 to 5.5 several weeks ago. During this upgrade I also changed the FQDN of the VCSA. I checked the used certificate and I was right. It used the old one. So I had to regenerate the existing certificate. Here is a blog post from William Lam (@lamw) if you will do it in an automated way.

To regenerate the certificate go to https://FQDN_VCSA:5480 and login with the root user. Then go to the Admin tab and select the following

After you select YES you have to reboot the VCSA.

During the reboot you will see that the appliance is regenerating the self-signed certificate.

Once the VCSA is fully booted check if the Hardware tab is back again.

Additional information about location of vCenter log files can be found here.

2,111 total views, 4 views today

Today I tried to install the new NexentaStor 4.0.1 on my x3650 M2 with local harddisks. To get the best performance my plan was to pass-through all local disks to the VM. For that I had to remove all RAID arrays that are currently configured across my 8 internal drives. After deleting the RAID arrays I started the host. When my host was up and running I tried to format one of the disks an got the following error message.

Call “HostDatastoreSystem.QueryVmfsDatastoreCreateOptions” for object “ha-datastoresystem” on ESXi “xxx.xxx.xxx.xx″ failed.

I have done this with several other disks and it looked like that only 3 had the problem. I enabled SSH and connected to the host.

There i used this command to list all the available devices:

ls -lha /vmfs/devices/disks/

This was the shortened output

As you can see there is a partition on the disks I couldn’t format.

The following command had shown me some information about the device.

partedUtil getptbl /vmfs/devices/disks/naa.5000c50023397a33

This was the output:

Error: The primary GPT table states that the backup GPT is located beyond the end of disk. This may happen if the disk has shrunk or partition table is corrupted. Fix, by writing backup table at the end? This will also fix the last usable sector appropriately as per the new reduced size. diskPath (/dev/disks/naa.5000c50023397a33) diskSize (286748000) AlternateLBA (570310655) LastUsableLBA (570310622)
Warning: The available space to /dev/disks/naa.5000c50023397a33 appears to have shrunk. This may happen if the disk size has reduced. The space has been reduced by (283562656 blocks). You can fix the GPT to correct the available space or continue with the current settings ? This will also move the backup table at the end if it is not at the end already. diskSize (286748000) AlternateLBA (570310655) LastUsableLBA (570310622) NewLastUsableLBA (286747966)
Error: Can’t have a partition outside the disk!
Unable to read partition table for device /vmfs/devices/disks/naa.5000c50023397a33

It looks like that after deleting the RAID array not all VMFS partition table information was deleted too, so I had to clear it manually.

To do this simply create a normal msdos partition on the affected disks with this command:

partedUtil setptbl /vmfs/devices/disks/naa.5000c50023397a33 msdos

After setting a new partition I was able to format the disk with VMFS5.

2,533 total views, 11 views today

When I wanted to convert an existing template back to a Virtual Machine this entry was greyed out. Also the entry for “remove from inventory” was greyed out.

Other VMs can be converted to templates and back. I found a Community thread were a User deployed a new VM from the template which can’t be converted and after that everything was fine again.

I have done the same an lo and behold after the deployment I can convert the template again to a VM.

But do I have to do this with every template? If you can restart the vCenter Server services than do this and the other templates will be fixed. If you can’t restart the service you have to deploy a VM for every template you cannot convert back to a VM.

2,430 total views, 15 views today

Today I’m very pleased to announce that after a couple of years Austria has again its own VMUG local group. The new local group leader is Andreas Lesslhumer (@lessi001, www.running-system.com) and I’m very honored to be a co-lead for this group. In the next couple of weeks we will sort out how many members are out there and how many of them are still active to form our first VMUG meeting.

About VMUG (VMware User Group)

Officially launched in August of 2010, the VMware User Group (VMUG) is an independent, global, customer-led organization, created to maximize members’ use of VMware and partner solutions through knowledge sharing, training, collaboration, and events. With over 90,000 members worldwide, we are the largest organization for virtualization users.

Our standing partnership with VMware has allowed us to create an ever-growing network of customers and partners who continue to strategically impact VMware products and services.

If you wanted to know more about VMUG please visit www.vmug.com.

German

Es freut mich heute bekannt geben zu dürfen, dass es nach längerer Zeit Abwesenheit, wieder in Österreich eine lokale VMUG Gruppe gibt. Der neue Leader dieser Gruppe ist Andreas Lesslhumer (@lessi001, www.running-system.com) und meine Wenigkeit die als Co-Lead diese neue Gruppe unterstützt. In den nächsten Wochen werden wir versuchen herauszufinden wieviele VMUG Mitglieder es in Österreich gibt und welche davon noch aktiv sind. Wenn wir dann eine ungefähre Zahl haben, machen wir uns daran unser erster VMUG Meeting zu organisieren.

Über VMUG (VMware User Group)

Offiziell gegründet im August 2010, ist die VMware User Group eine unabhängige, globale von Endkunden geführte Organisation, die das Ziel hat Wissen über VMware als auch Partner Produkte durch Gedankenaustausch, Schulungen, Zusammenarbeit und Events zu maximieren.

Mit weltweit über 90.000 Mitgliedern ist Sie die größte Organisation für Virtualisierungsbegeisterte.

Die feste Partnerschaft mit VMware hat es uns erlaubt ein ständig wachsendes Netzwerk von Endkunden und Partner zu schaffen, welche ständig strategische Auswirkungen auf VMware Produkte und Services haben.

Wenn Sie mehr über die VMUG erfahren wollen besuchen Sie bitte www.vmug.com.

1,658 total views, 7 views today

A week ago I was at a customer site to upgrade from vSphere 5.1 to vSphere 5.5 U1a. The customer has IBM Rack Server and is using Cisco Nexus 1000v. First I tried to get an customized image from IBM but the only available was for 5.5 RTM. Another problem is that all of this images doesn’t include Nexus 1000v VEMs.

For this case I used vSphere PowerCLI Image Builder cmdlets to build a customized images with VEM, IBM CIM providers and updated drivers from the VMware site which are also included in the IBM customized image.

First I created a folder for all the offline bundles and the most recent VMware ESXi offline bundle.

When going to https://my.vmware.com/group/vmware/patch#search and search for e.g. ESXi and 5.5 the top package is the most recent offline bundle.

1. Create software depot

To create your ESXi software depot run following command:

Add-EsxSoftwareDepot Most_Recent_ESXi_offline_Bundle

Add-EsxSoftwareDepot ESXi550-201404001.zip

Leverage this command to include all of your need offline bundles to the software depot.

Add-EsxSoftwareDepot BCD-bfa-3.2.4.0-00000-offline_bundle-1634864.zip

Add-EsxSoftwareDepot BCD-bna-3.2.4.0-00000-offline_bundle-1634921.zip

Add-EsxSoftwareDepot ibm_sw_hyper_provbundlv1_vmwaresx55_32-64.zip

Add-EsxSoftwareDepot igb-5.2.5-offline_bundle-1682588.zip

Add-EsxSoftwareDepot ixgbe-3.21.4-offline_bundle-1710123.zip

Add-EsxSoftwareDepot megaraid_sas-6.603.55.00-offline_bundle-1712343.zip

Add-EsxSoftwareDepot qla4xxx-634.55.28.0-offline_bundle-1682448.zip

Add-EsxSoftwareDepot qlcnic-esx55-5.5.183-offline_bundle-1666026.zip

Add-EsxSoftwareDepot VEM550-201308160108-BG-release.zip

Add-EsxSoftwareDepot VMW-ESX-5.5.0-qlnativefc-1.1.20.0-offline_bundle-1604804.zip

If you don’t want to download the ESXi offline bundle prior to the command you can use

Add-EsxSoftwareDepot https://hostupdate.vmware.com/software/VUM/PRODUCTION/main/vmw-depot-index.xml

to include the whole online depot.

2. Get most recent Image Profile

Using

Get-EsxImageProfile

you will get all Image Profiles which are in your software depot. You will see a -standard and a -no-tools profile. I would recommend to use the -standard profile because it includes VMware Tools.

When using the online depot you will get a bunch of offline ESXi bundles. Use

Get-EsxImageProfile | Sort-Object “ModifiedTime” -Descending | format-table -property Name,CreationTime

to get the most recent on top of the list. It could happen that there are still a lot of bundles with the same ModifiedTime. In this case you have to check https://my.vmware.com/group/vmware/patch#search for the most recent one.

3. Clone most recent Image Profile to new Image Profile

To create a new Image Profile simple clone it using

New-EsxImageProfile -CloneProfile MostRecentImage “ NameOfNewImage” -vendor NameOfPartner/Vendor

New-EsxImageProfile -CloneProfile ESXi-5.5.0-20140404001-standard “ESXi-5.5.0-20140404001-standard-IBM-VEM” -vendor Custom

4. Show all Software packages in the software depot

To get a list of all available software packages in the created software depot you can use

Get-EsxSoftwarePackage

5. Add required software packages to the Image Profile

Software packages with VMware as vendor name are automatically included in the copied Image Profile. To include all other necessary packages use

Add-EsxSoftwarePackage -ImageProfile Name_of_copied_Image_Profile Name_of_software_package

I used the following commands to add IBM CIM provides, most recent drivers and Cisco VEM module to the Image Profile

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM cisco-vem-v160-esx

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM scsi-megaraid-sas

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM net-igb

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM net-ixgbe

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM scsi-qla4xxx

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM net-qlcnic

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM qlnativefc

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM ima-qla4xxx

These are the IBM CIM provides which should added in the following order because e.g. filetrans depends on concretejob. If you add filetrans before concretejob you get an error message that a dependency is missing.

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM hwckvm

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM fupb

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM pciinfo

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM soibms

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM immpassthru

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM concretejob

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM filetrans

Add-EsxSoftwarePackage -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM fwupdate

6. Create a new ESXi Install ISO based on the Image Profile

After all software packages were added to the Image Profile you can create an ISO file based on the Image Profile

Export-EsxImageProfile -ImageProfile Image_Profile_Name -ExportToIso Name_of_new_ISO.iso

Export-EsxImageProfile -ImageProfile ESXi-5.5.0-20140404001-standard-IBM-VEM -ExportToIso ESXi-5.5.0-20140404001-standard-IBM-VEM.iso

You can now use this newly created ISO with vCenter Update Manager to upgrade existing ESXi servers.

1,344 total views, 11 views today

Today I was at a customer site and the customer told my a really strange story and I don’t know if everybody else is/was aware of the following behavior. But first of all the story.

He was contacted because a Terminalserver was no longer reachable by the IP address. He looked at  the VM settings and found out that the NIC was gone. He added a NIC back to the VM and the Terminalserver was reachable again. This happened 2 weeks ago and he told me the problem today. So this story immediately awakened my curiosity and troubleshooting ambition.

First I checked Task/Events to see if there was a manually removal within the vSphere Client. but I found nothing. The only thing I found was the entry where the customer added back the missing NIC. After that I downloaded the vmware-##.log from that VM at a given period and found the following:

Powering off Ethernet0
Hot removal done.

My first thought was “OK….. and who did that?”. I did some research and found this VMware Blog post. The following information is taken from this post.

There are 2 reasons why a NIC go missing from the VM.

• Hot Removal from the Guest. With the new Hot Add/Remove feature, NICs show up under the “Safely Remove Hardware” list. Any user with administrative privileges can accidentally remove the NIC using this feature.

• Someone manually removed it from the Virtual Machine configuration (Probably using UI or some SDK APIs).

If somebody removes the NIC through vSphere Client or SDK APIs you will find the following in the vmware-##.log within the Virtual Machine folder:

Requesting hot-remove of ethernet1

If the NIC is removed within the guest OS using “Safely Remove Hardware” you will see the messages mentioned earlier. What I didn’t know was the fact, that if the NIC is  removed in that way it is also removed from the VM settings! I thought until today that this function will only remove the NIC from the guest OS and will be back after a reboot.

There are 2 ways to avoid this problem.

1. Disable the HotPlug capability of the VM

If you disable HotPlug of the entire VM you can’t add HDD and NICs if the VM is powered-on.

1. Connect to the ESXi/ESX host or vCenter Server using the vSphere Client.
2. Power off the virtual machine.
3. Right-click the virtual machine and click Edit Settings.
4. Click the Options tab.
5. Click General > Configuration Parameters > Add Row.
6. Insert a new row with the name devices.hotplug and a value of false.
7. Power on the virtual machine.

When you try to add an hard disk or network card you will get this:

2. Use Windows Registry hack to hide the hot removeable capabilities of the NIC

The post I mentioned previous is from 2010 therefor some procedure are outdated for Windows 2008 R2/2012/2012 R2. To edit the desired value you have to start the registry editor as Local System. To do so download either the complete Sysinternals Suite or only the psexec Tool from the Microsoft Sysinternals site.

1. Open a Command Line Prompt (cmd)
2. Execute psexec -i -s regedit.exe
3. Go to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum and search for your NIC driver (VMXNET3, E1000 etc.)
4. Set the Capabilities flag in the key(s) found above, to the current value minus 4

For example I have the key HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\PCI\VEN_15AD&DEV_ 07B0&SUBSYS_07B015AD&REV_01\FF5650006DFF91FE00 (1) with the Service value (2) vmxnet3ndis6. The capabilities value is set to 16 (3). On changing the value to 12, (immediately) VMXNET3 NIC will be no longer listed in the safely remove hardware list anymore.

After Capabilities changed to 12

NIC is hidden from “Safely Remove Hardware”

774 total views, 7 views today

I was attending the Veeam Certified Engineer course this week and tried some things in my home lab. One of the discussions in the course was, if it’s possible to add a ESXi Hypervisor to Veeam Backup & Replication. From my view it should be possible to add it and use it as a replication target.

I deployed an ESXi from a template, add a datastore for the replicas and added it to the VBR server. Afterwards I created a Replication Job to replicate VM-1 from my production ESX to the ESXi Hypervisor (Free license).

This job failed with the following error message:

Processing configuration Error: Current license or ESXi version prohibits execution of the requested operation.

So you can add a vSphere Hypervisor host to the VBR server but you can’t use it for Backup or Replication because you need the vStorage API.

I decided to delete the Replication Job and the standalone host but I got the following error when I tried to delete the host after the job was deleted:

That was interesting because I deleted the job so there shouldn’t be a job anymore. So the only way to get rid of this “ghost” job was to edit the Veeam B&R database.

I would think that the following procedure is not supported by Veeam.

Open SQL Management Studio and log into the server

Go under Veeam Database – Tables

Right click Backup.Model.Backups and select “Select Top 1000 rows“, there you will find the missing Job with a job_id of NULL

If you try to delete this row you get an error which states that there is a REFERENCE contraint in the table Backup.Model.Points.

Before you open this table write down the id of the missing job

Delete in the table Backup.Model.Points the row with the matching backup_id value

After you have deleted both rows you can delete the standalone host within the VBR server.

To verify that this will only happen when you use the free license I put an Essentials license in the ESXi host. I created the Replication Job and it successfully completed.

Then I deleted the Replication Job and tried to delete the standalone host but I got the same error message as with the free license. I searched for the replicated VM in the VBR server, right clicked on it and selected “Remove from disk”. After this procedure I was able to delete the standalone host.

Looks like that there is a little bug in Veeam Backup & Replication 7 Update 3 when you try to use a free licensed ESXi host.

422 total views, 7 views today

Last time a colleague came to me and ask me how to pass-through a license dongle to a VM. The first thing that came to my mind was “Hm…. haven’t done that before, but should not be so complicated”. I tried it in my home lab and this post is the result.

First of all make sure that the device you wanted to use is supported for pass-through. To validate it check this VMware KB article.

To use USB pass-through the VM must have Virtual Hardware Version 7 or later.

Methode 1

Open the WebClient, select the VM which will use the USB device and click Edit Settings.

Select under New device Host USB device and click add.

This will also add an new USB controller as you can see in the screenshot. If you have more than one USB device attached to the ESX you can select the appropriate one through the drop down menu.

When you switch back to the VM you can see the installation of the new components.

After everything is installed, you can find the device within the device manager or if it’s an USB stick, like in my example, under Computer.

Methode 2

There is also an alternative method, but I would not recommend this. You can also pass-through a whole USB Controller from the ESX.

The use a physical USB controller select the ESX in the Webclient and go to Manage – Settings – PCI Devices

Select Edit… and add the desired USB controller.

After you added the controller you have to restart the ESX server.

You can now add the PCI device (USB controller) through Edit Settings – Add Device – PCI device to the VM.

ATTENTION

You can’t disable the USB controller after that methode when you’re using USB sticks or SD cards to boot your ESX server. This issue occurs when the image is loaded to the memory which includes the esx.conf file. When you enable the USB controller for pass-through, all USB devices are invisible to the kernel. When you deselect the USB controllers from the pass-through list, the change saves to the esx.conf file. The esx.conf file is in the memory during ESXi uptime and when you issue the reboot command, the esx.conf is saved back to the USB device. In this case, the kernel cannot see any USB devices, so the configuration change does not apply.

The only option to disable the pass-through controller is to do a fresh installation of the ESX server according to this KB article.

373 total views, 12 views today

With VMworld EMEA 2 months away and VMworld US kicking in in 1 week I thought I can give some Tips/Tricks/advice that I’ve learned over my last 5 VMworlds. Some of these are general for both, US and EMEA conference and some are especially for EMEA conference.

General Tips & Tricks

Wear comfortable shoes and clothes – my preference for dress code is Jeans, T-Shirt/Polo and sneakers.

Use large luggage – when you visit the Solution Exchange you will get a lot of vendor swag especially T-shirts you have to carry home!

Attend Partner day – there is a special day for partners where you get the latest info about the partner program, how to sell products, acquiring new customers, etc.

Get certified – If you want to get certified with VCP or VCAP use the 50% off promotion when you are at VMworld. This can save you 200$ when doing a VCAP. I’m using the partner day for exam scheduling so I don’t miss the other cool stuff.

Schedule your session - When scheduling your sessions do not schedule to many sessions. You will have long walks between the sessions especially at the EMEA conference and you will do other stuff beside sessions like visiting the Solution Exchange, doing some labs or get in touch with other people.

Network, Network, Network – at my first VMworld I attended a lot of sessions because I didn’t know anybody. Since I have started blogging and getting in touch with all the great people, the sessions I schedule become less and my time I spend in the Solutions Exchange and at the hang space become more and more

Attend Parties – there are a couple of vendor parties everyday. The best I was at were Veeam, PernixData, Nutanix and certainly VMworld Party itself. If you’re a vExpert or VCDX you should also attend vExpert/VCDX party.

Taking pictures – at one VMworld I carried my DSLR with me. With the backpack (incl. Laptop) and the DSRL bag it’s getting very heavy over time. So use a small camera – it should be sufficient even when you take pictures of the general sessions.

Get a Twitter account - it’s easier to get in touch and follow people. You will also get a lot of information when using twitter e.g. new blog posts or information about parties at VMworld.

EMEA conference Tips & Tricks

Watch your wallet – In Barcelona there are a lot of pickpockets so don’t put your wallet in the back pocket or it may get stolen.

Free time – if you have free time, take the opportunity to visit the most interesting spots in Barcelona (Park Guell, Sagrada Familia, Magic Fountain of Montjuïc, Las Ramblas, Gothic quarter). One personal recommendation if you want to buy fresh fruits and fresh fruit juices visit the La Boqueria market, it’s quite in the middle of the Las Ramblas.

Public transportation – If it’s like the last years you will get a public transportation pass with 10 rides (sponsored by Veeam in the last years). You will get to the conference center (Fira Barcelona Gran Via) when using the lines L8, S4, S8 or S33 from Placa Espanya to Europa Fira station. From the station there will be a shuttle bus service to the conference center. Don’t try to walk, it will take 20-30 minutes because it’s on the north entrance of the Fira.

Airport transfer – There are shuttle buses which goes directly from the BCN Airport to the conference center. Search for the ladies with the VMworld sign!

Attend vRockStar Party – if you’re in Barcelona on Sunday then you have to attend vRockstar party invented by Marco Broeken and Patrick Redknap. Follow these guys on twitter to get the latest updates about the party.

Carry a vest/light jacket with you – October in Barcelona is quite warm (20-24°C) but the conference facilities are all air conditioned and it’s getting a little bit cold inside. If you will take an exam DON’T take a vest/jacket with you because you have to take it off for the exam (security guidelines). Use a sweater or something similar instead. See my last year’s exam experience here.

More Information

If you’re not registered for VMworld yet then use this link to do so and get involved in the most awesome conference of the year.

For more information about VMworld follow this link.

238 total views, 11 views today

I had this problem several times in the last few months. In a small environment the customer is using SQL 2008 R2 Express Edition for their vCenter database. When using SQL 2008 R2 Express you must be aware of the following limiations.

SQL 2008 R2 Express limitations (taken from here)

Because the memory limit is only applicable for the buffer pool, the SQL process can take more than 1GB memory in total because not all memory allocations go via the buffer pool.

After a while the customer observes the following message in the Application event log:

The transaction log for database “<vCenter database name>” is full. To find out why space in the log cannot be reused, see the log_reuse_wait_desc column in sys.databases.

This indicates that the SQL database is almost at it’s limit of 10GB. To find out why the database grows that much I used a SQL script I found here.

In SQL Server, the page size is 8 KB, or 128 pages per megabyte
*/ 

select
    a2.name [Table],
    replace(convert (nvarchar, convert(money, a1.rows), 1), '.00','') [Row Count],
    cast (a1.reserved/128.0/1024 as decimal(9,3)) [Reserved GB],
    case when a1.reserved = 0 then 0
    else cast ((100-(100.0*a1.used/a1.reserved)) as decimal(9,2)) end [Unused %]
from (
    select
        ps.object_id,
        sum (cast (ps.row_count as bigint)) rows,
        sum (cast (ps.reserved_page_count as bigint)) reserved,
        sum (cast (ps.used_page_count as bigint)) used
    from sys.dm_db_partition_stats ps
    inner join sys.all_objects a2
		on a2.object_id = ps.object_id
		and a2.type IN ('U')
    group by ps.object_id
    )  a1

left join sys.all_objects a2 ON a1.object_id = a2.object_id
where (a1.reserved/128) > 5
ORDER BY a1.rows DESC, a2.name

When running this statement against the vCenter database I got the following result:

As you can see there are 2 tables which are quite big and has a lot of rows. This will happen when you leave the following with default settings:

If something in your environment creates a lot of tasks and events (e.g. using Nagios monitoring) then both tables will grow bigger and bigger. To avoid this activate task and events clean-up in the vCenter settings and set an appropriate retention time. My customer used 30 days for it.

But how can I shrink the current tables? You have 2 options for that.

1. Wait until the clean-up stored procedure runs

2. Run the clean-up stored procedure manually

If the situation is urgent, which means vCenter is down because the database is full, connect to the database server through SQL Server Management Studio and go to vCenterDatabase > Programmability > Stored Procedures, right-click dbo.cleanup_events_tasks_proc and select Execute Stored Procedure.

It will take some time until the clean-up is finished, depending on the numbers of rows you have in both tables.

After the clean-up and executing the SQL script you will have a similar output:

For more information visit the original KB article.

97 total views, 11 views today

Since 2005 I have been employed by VMware solution & service provider partners and I have learned a lot about all the different VMware products. Over time virtualization especially with VMware took more and more part in my daily business and also my career path has changed over the years.

I started as a Systems Engineer (for Austrian companies it is similar to an Implementation Engineer) in a small company doing a lot of Windows, Exchange, network and since 2007 also small virtualization deployments starting with ESX 3.5. When I started working for one of the biggest VMware partners in Austria, I began to get more in-depth understanding and knowledge about the products VMware offered at this time. In my current position as a Senior System Consultant at Imtech ICT Austria GmbH I started my VCDX journey to advance myself in terms of design aspects. Since 2010 my daily work also consisted of pre-sales and post-sales consulting including infrastructure designs for customers. In this year I decided that I wanted to work for VMware in the future.

I had my first opportunity in 2011, when I applied for a Senior Consultant position. Unfortunately my lack of experience has cost me this opportunity. So I studied harder, got more experience, did a lot of VCP and VCAPs and had 2013 another chance to join VMware as a Senior Consultant for Automation. Again my lack of experience and knowledge in a particular product/topic cost me the position. After that I was really upset and thought I would never work for VMware.

After all these downs exciting times lay ahead: I’m really happy to take on a new challenge and finally join VMware as a Technical Account Manager, starting on Monday (1st of September 2014).

For all who are not familiar with a Technical Account Manager, in short TAM, here is a short overview:

To be honest I’m so excited that I can’t wait to start working for my new company. One of the best things is that I will work again with my former mentor Henrik Andersson (@vmswede) who also will be my mentor at VMware.

39 total views, 7 views today

Yesterday Henrik Andersson and I discovered a strange problem. The following screenshot shows the problem we saw in our demo environment. To be clear, this error showed up as soon as we clicked Create new VM Storage Policy.

com.vmware.vim.binding.vmodl.fault.SystemError:

reason = PbmExtendedElementDescription required property summary not set

inherited from com.vmware.vim.binding.vmodl.fault.SystemError: Failed to serialize response

We went trough the troubleshooting process and found the one or other hint which can lead to the problem.

One of the hints was that when using Tags or Tags categories with no description you can encounter this problem. We had 4 Tags in our environment and after entering a description for all 4 the problem was gone.

To be sure that this was the solution I tested it also in my home lab. I created 1 Tag with a category and started the Create new VM Storage Policy wizard.

To my regret the error didn’t show up. So what’s the different between my home lab and the demo lab. They are different in ESX build version, enabled features (demo lab has VSAN enabled) etc.

We tried to delete all existing Tags and recreated it. Henrik created the Tags with the following Powershell cmdlet.

New-Tag -Name "TestTag" -Category "TestCategory"

I created my Tags manually. When testing the behavior again, my wizard shows no error but Henrik wizard did. I told him to create the Tags also manually and after that the error was also gone.

There is obviously a problem when creating Tags with the Powershell cmdlet and not using the discription parameter. We think that the Powershell cmdlet is setting a NULL or similar value for the Tag description and therefor the error occurred.

If you have any further information about that, please leave a comment.

117 total views, 2 views today

I’m very happy that I can attend my 6th VMworld in a row and I’m very pleased and honored that VMware selected me for a blogger pass. I will do my best to spread as much info as possible during the conference. Over the years I learned a lot about how to attend VMworld. For this year my primary focus will be the hang space and the Solution Exchange. I have also scheduled some session I’m interested in, but I can watch them after VMworld so there is no need to attend to much. Why the hang space? It’s a place where I can meet people (twitter follower, blogger fellows), I can’t get in touch  personally over the year because they are spread across the globe. I’m really looking forward to some great tech talk with all the awesome guys and ladies! It’s also a great place to write some post if time will permit it.

The second area you will find me is the Solution Exchange. I love to learn about other technologies beside VMware and the Solution Exchange is the place where you get all the information and product demos of the technology you’re interested in or maybe you don’t even know that such product exists.

My VMworld Sessions

Here is a list of sessions I will attend this year.

VMworld Gatherings

After a long day at the conference it’s time to relax at one or more of the gathering that happens during VMworld. Here is a list of all gathering so far. If you need an updated list, please click here. You can also see which activities I will attend.

VMworld vBrain.info Swag

For the first time, since I started blogging, I will also have some swag in my bag. Andreas Lesslhumer (@lessi001) and Chris Wahl (@ChrisWahl) gave me the idea of creating my own buttons for VMworld. If anybody wants some, find me and get one. I hope I will spread half of the 100 buttons I got with me!

I will be accompanied by my friend Andreas Lesslhumer (first vExpert in Austria, VMUG Leader Austria) who also has some buttons to spread. If you will see us don’t hesitate to join us for a little tech talk if you want.

Transportation: Airport and Metro Shuttles

A shuttle service is offered to attendees to/from the airport.  Attendees will be greeted in the arrival hall. Look for the VMworld 2014 sign on arrival. There is no shuttle service offered to/from hotels.

Airport Shuttles
Shuttle drop off/pick-up: Fira Gran Via, North Entrance, Hall 8.0.
Sunday, 12 October: 16:00-20:00
Monday, 13 October: 10:00-20:00
Thursday, 16 October: 12:00-18:00

Metro Shuttles
Shuttle drop off/pick up: Fira Gran Via, North Entrance, Hall 8.0.
Sunday, 12 October: 14:00 – 20:00
Monday, 13 October: 07:00 – 21:00
Tuesday, 14 October: 07:00 – 20:00
Wednesday, 15 October: 07:00 – 00:00
Thursday, 16 October: 08:00 – 16:00

Safety in Barcelona
Cause of my 3rd VMworld in Barcelona it would be advisable to remove your VMworld badge when you are out enjoying Barcelona, so you don’t draw unwanted attention to yourself as a visitor. Same should count for the VMworld back bag, although I didn’t have any problems with that in the last years.

If you need further Tips & Tricks for your first VMworld or your first in Barcelona you can check out this post.

Finally I would like us to have an awesome VMworld 2014.

81 total views, 8 views today

At VMworld US there is always a traditional vBreakfast at Mel’s Diner invented by Shane Williford. So Jane Rimmer, Pietro Piutti and Andres Prieto discussed on twitter that there should also be a #vBreakfast for VMworld Europe. Andres and Pietro thankfully done some research for possible locations near the Fira and came up with this location

Cafeteria Praga
Carrer l’Energia 24
5min walk from Fira2
Map

What better way to get to know the community than over breakfast. Join me and many others in this hopefully getting yearly tradition of carbo-loading for an action-packed day on Tuesday, Wednesday and Thursday starting at 7:45.

For more information follow Jane, Pietro and Andres and watch out for #vBreakfast and #vBreakfastBCN.

See you there!

18 total views, 18 views today

I arrived at Barcelona with some delay because there was a lot of traffic up in the air. Also this “traffic jam” causes a much longer delay at the luggage retrieval because it started 45 minutes after our arrival. A friend of mine got stuck at the Frankfurt airport and arrived in BCN 8 hours later than expected. Beside all of the problem we had at the beginning we have great weather in Barcelona but humidity is killing me, especially in the Metro stations. After some lunch with my Austrian friends we headed to the Fira to get registered for an awesome week of tech and talk.

After registration was finished we went back to our hotel to get some rest before one of the best parties at VMworld kicked in.

vRockstar party is the unofficial pre-conference party organized by some really great people (Patrick Redknap, Marco Broeken, Robert Payne & Hans De Leenheer). Like last year it took place at the Hardrock Café, Plaça de Catalunya, 21 and was sponsored by Cisco, EMC, Nutanix, VMUG. This year the whole main floor was reserved for the party which was really awesome. The only thing what was not so optimal was the temperature and the noise level. If you wanted to talk to each other you had to get out of Hardrock Cafe or shout at people.

Here some pictures from the party.

14 total views, 14 views today

After a really awesome vRockstar Party on Sunday, VMworld was official started with the Partner Day for VMware Partners and also the TAM (Technical Account Manager) Day for TAM customers. We arrived at 8:30 at the conference center and tried to get some breakfast there. Unfortunately there was only a small Cafe within the Fira where we had to pay for our breakfast. I remembered that at the last Partner Day in 2013 there was breakfast for the Partners. Also none of the beverage boxes where opened at that point. Little bit upset I joined the VMUG HOL VIP Tour I had registered for.

The VIP Tour was really good because you got a little bit of inside what was happening in the Hand on Labs during the whole conference. The HOL where all hosted in either the Amsterdam or Santa Clara datacenter. The Amsterdam datacenter was build of legacy servers and storage. According to the HOL architects the whole infrastructure was backed by a 5 rack storage infrastructure in the last years. Now ONLY! a 12U EMC XtremeIO box was used! What an impressive change of consolidation. The datacenter in Santa Clara was build upon EVO:RACK. If you don’t know what EVO:RACK is please visit this site for more information. The complete infrastructure was monitored by vRealize Operations Manager and vRealize Log Insight. I was told that over the complete conference Log Insight would gather 2 billion entries. We saw during the Tour that there were already 500 million entries captured. The deployment process is done via vCloud Director. The self service portal where you can request a specific Lab is a self-developed application which properly could be available for purchase in the near future.

After the tour I had a cool conversation with Joe Silvagi (@VMPrime) about his VCDX defence which was a week before the conference in Palo Alto.

On of the best moments at this day was when somebody came over to me because he saw my button on my friends backpack, recognised the blog on it and told me he is a regular reader of it. These are the moments I realise that my hard work pays off.

Going back to the registration desks I found out why there was all beverage boxes locked. I missed the Partner Day area between the registration desks! I tried to get in but was rejected. A lovely lady told me that when attending the conference with a blogger pass you can’t join the Partner Day session, even though you are a partner, because it could be that there is some NDA information, VMware wouldn’t want you to blog about. Hey when I’m a partner I already have signed an NDA agreement. Maybe there can be an update on the blogger pass for next year.

The rest of the day I spend at the hang space doing some bog posts etc. On 4pm I joined the EMC/Avnet party at the Camp Nou. It was really cool because I had an Tour through the whole stadium.

Camp Nou Stadium Tour

I left the party at 8:30 because I was also registered for the VMUG members party as well as for the PernixData Party. PernixData party took place at the Ocana club and it was a really cool location. There I had great conversations with Gabriel Chapman (@Bacon_is_King, Amy Lewis (@Commsninja) and Frank Denneman (@FrankDenneman) and it was really funny to watch the teasing between Amy and Frank. One of the best quotes according to my “Frank! Amy will be one of your coffin nails” was “One? No! All! I also get the chance to see how this “awesome” photos of Josh Atwell (@Josh_Atwell) were created and all started with a BAD IDEA.

What happens at the pernix party is immediately distributed. #notvegas // @Josh_Atwell @FrankDenneman pic.twitter.com/RPSCXYUiCG

— Amy Lewis (@CommsNinja) October 13, 2014

In the end I also had the chance to talk to Erik Sloof (@esloof), Marek Zdrojewski (@MarekDotZ), Pietro Piutti (@stingray92) and Francesco Bonetti (@fbonez).

After an awesome day we got back to the hotel at 1:30am to get same sleep before we started the next day with a vBreakfast.

105 total views, 3 views today

My day at VMworld Day 2 started at 6:30 because we (Austrian “Gang”) wanted to participate in the vBreakfast. We found the Cafe near the Fira on time. Unfortunately we were the only one there. A fourth guy, Fabian Lenz (@lenzker), a VCI from Germany joined us for breakfast but that’s it. Jane and Pietro tweeted me that they can’t make it. I took this first approach as a beta test and hoped that there will be more people on Wednesday and Thursday. To be honest I hadn’t checked this because i didn’t attend the other vBreakfasts because of all the parties that was going on in the evening. Overall there is a lot of room for improvement for the next years vBreakfast at VMworld Europe.

The good thing about being up this early was that we were not late for the General Session. I don’t want to summaries the General Session and all the announcements because it was already done by some fellow bloggers like Magnus Andersson, Alex Galbraith and Michael Poore. This Info could be found here:

Magnus Andersson – VMware Barcelona 2014 – Day 1 Summary

Alex Galbraith – VMworld Europe 2014 – Keynote Day One

Michael Poore – VMworld 2014 – Day 1

Keynote VMworld Europe 2014 Day 1

After the General Session we headed to the blogger desk to spread some of our buttons there. Andreas had really cool buttons for this VMworld. My personal goal was to spread approx. 50% of them. At the blogger desk we found also some other cool blogger swag lying around. A Hyper-VCDX and blogger sticker by Joup Piscaer (@jpiscaer), a beer coaster and cool sticker by VMguru.com (@vmguru).

After lunch (was really good compared to the US conference) I went to the Solutions Exchange which was my primary place to be this VMworld, because I wanted to talk to several Vendors about their solutions. Because I joined VMware doesn’t mean that I lost completely my independence. I’m still interested in other technology that can benefit my customers in some areas. My first visit was the VMware expert bar to meet finally Frank Buechsel (@fbuechsel) in person. He is an cool and awesome guy and a challenge when it came to certifications. There is always some competition between him, Ross Wynne (@rosswynne) and myself when it comes to VMware certification.

My first stop on my Vendor tour was Veeam for a quick chat with Claus Pfleger (Systems Engineer Austria) and Phillip Altmann (Territory Manager East Austria) about the new “free” product (Veeam Endpoint Backup) what was announced earlier that week.

Diablo Technologies who contacted me previous to VMworld for a chat/demo was my next stop because I had no time for this demo so I told them I will stop by their booth at the conference. Their product, Memory Channel Storage (MCS), is very impressive because data capacity is now accessible via the shortest and fastest possible path to the CPU, the memory channel. This is enable by a driver you need to install on the ESXi host. There is no prerequisites for the server (special Server, BIOS etc.) to use this DIMMs, as long as the specification for memory in the server is DDR3. The only thing to keep in mind about this technology is, that you will lose memory capacity when using MCS. For more information about Diablo Technologies visit their website.

Another storage vendor I took a look at was SolidFire. Solidfire is a scale-out All-Flash array. It uses a node-based scale-out architecture, that means if you run out of capacity or performance you can add nodes to get more of both. When adding a new node(s) Solidfire rebalance non-disruptive the capacity and performance in minutes. A cool thing is also that you can also remove nodes from the system and create new clusters within another location. The minimum for a cluster is 3 nodes but Solidfire recommends 4-5 nodes. Another key feature of Solidfire is the Guaranteed Performance QoS, that means you can create a Volume with a min, max and burst limit for IOPS and you can change this on the fly. Solidfire uses Shared Nothing High Availability which mean a cluster-wide RAID-less data-protection called SolidFire Helix spreads at least two redundant copies of data across all drives within the system. There are also more features like In-line Efficiency or Automated Management. If you wanted to know more about SolidFire please visit their website.

On my way to the PernixData booth I ran into Erik Bussink (@erikbussink). We talked about our homelabs and how they evolved over the years. We also talked about the new job and how each other came to his position. It was a great chat and Erik is really an awesome guy to talk with. Erik reminded me also that I would get a PernixPro jacket but my size was unfortunately out of stock. PernixData told me that they will ship it to me after the conference (thanks God! :)).

The first day was really crazy because I couldn’t walk 5 meters without running in some familiar fellow blogger or twitter follower like Pietro Piutti (@stingray92) or José Luis Gómez (@pipoe2h)

I joined Andreas again after some time and he told me he got a really cool T-Shirt from Atlassian (@atlassian). -> The ANGRY NERDS Shirt. That was by far the best T-Shirt at VMworld Europe 2014. For people like me who didn’t know who Atlasssian is and what they do here is a brief overview.

We came across PernixData booth where I meet their CTO, Satyam Vaghani (@SatyamVaghani). He congratulated me about my shift to VMware. I had a great chat with him about PernixPro program and the new FVP 2.0. He is really a great guy and I hope I can retain my PernixPro status. although I change to VMware.

In the evening I was attending the vExpert/VCDX/NSX party where I finally meet Corey Romero (@vCommunityGuy) and for all of us who don’t know him, he is the vExpert Program Manager. It was really a pleasure talking to him about the program.

92 total views, 9 views today

From time to time I learn something new about features in vSphere and other products I thought I was familiar with, so I decided to create a series of blog posts under the topic “Things you should know” or in short TYSK. This series of “short” posts will be focused on topics/features etc. everybody is using in his daily business/job but forgets over time some of the details.

Couple of days ago I had an discussion with a colleague about Cross-host vMotion and in which license it is included.

First let’s start with the basics! What is Cross-host vMotion? Cross-host vMotion was introduced in vSphere 5.1 and is the ability to change both host and datastore at the same time during a migration. It’s only available through the vSphere Web Client, where you can change both host and datastore of a running VM.

Testing procedure

For testing purpose I installed  a small environment with 2 nested ESXi hosts and created a cluster in my main vCenter. I created a VM within the nested environment to test Cross-host vMotion (Shared-nothing vMotion, vMotion without shared storage or however you will call it).

The following shows the requirements for Cross-Host vMotion (for more information click here):

• The hosts must be licensed for vMotion.
• The hosts must be running ESXi 5.1 or later.
• The hosts must meet the networking requirement for vMotion.
• The virtual machines must be properly configured for vMotion.
• The destination host must have access to the destination storage.

As you can see a license where vMotion is enable can be used for Cross-Host vMotion which means anything from Essentials Plus to Enterprise Plus can be used. To test it I installed a RoBo version of an Essentials Plus license on my nested Hosts and rebooted it to make sure that the license is correctly loaded.

After that I migrated the Test-VM from one host (Storage Local01) to the other host (Storage Local02).

As you can see in the screenshot Storage vMotion is not licensed.

In my nested environment it took some time because the VM had approx. 40GB. When Cross-Host vMotion was finished you can see that the host and datastore were changed.

I thought that this can also be used to migrate the VM from one datastore to another datastore without changing the host, but after I completed the migration steps I got this message:

Conclusion

When you will use Cross-Host vMotion, the only licensed feature you will need is vMotion which is included in Essentials Plus and higher.

76 total views, 3 views today

I made recently a really cool discovery with vStorage APIs for Array Integration in short VAAI, but first what is VAAI.

• Offload block copy and block zeroing to the array.
• Support dead space reclamation and out-of-space warnings in thin-provisioned arrays.

For more information about all the Primitives VAAI is supporting I would recommend this blog post by Chris Wahl and this post by Cormac Hogan.

As most of us know, VAAI is included in vSphere Enterprise Edition.

And now the discovery! It looks like that VAAI is also working on all other vSphere Editions starting from vSphere Essentials Plus.

Testing scenario

I’ve installed 2 nested ESXi hosts and used shared storage with VAAI capabilities. In my case it’s a HP StoreVirtual VSA which is supporting the following VAAI Primitives:

I created 2 VMs. One with a thin provisioned disk and one with a thick provisioned disk. For the test I was using an Essentials Plus RoBo license and an Enterprise license. Between all license changes I rebooted the ESXi hosts to make sure that the license and features are all correctly loaded.

VM (thin disk)	VM (thick disk)

As you can see in the following screenshots only the Enterprise license shows the availability of VAAI.

Essentials Plus

Enterprise

Testing procedure

During my tests I was connected through ssh (putty) to the ESXi hosts and had an instance of esxtop opened on each of them. In esxtop I selected u (disk device) and used only the fields A, B and O to display only VAAI relevant metrics.

My tests were the following:

• Clone powered off thin disk VM (Essentials Plus)
• Clone powered on thin disk VM (Essentials Plus)
• Clone powered off thick disk VM (Enterprise)
• Clone powered on thick disk VM (Enterprise)

My assumption with these tests was that when it will work with one disk format, it will also work with the other disk format when using the same license.

Clone powered-off thin disk VM (Essentials Plus)

You can see that the Clone Primitive values and the Zero Primitive values are going up, which means VAAI is used.

Clone powered-on thin disk VM (Essentials Plus)

Also on the powered-on VM VAAI values going up and VAAI is used.

Clone powered-off thick disk VM (Enterprise)

Same picture with the Enterprise license. Here it is expected that VAAI is used. In our case only the Clone Primitive is used because it’s a thick provisioned disk.

Clone powered-on thick disk VM (Enterprise)

Last but not least the final test showed also that VAAI is used which is expected when using Enterprise licensing.

Conclusion

After my tests it is clear that VAAI is always used regardless of vSphere Edition. When looking at the Product Guide, which can be found here, you will find on page 13 the following excerpt:

vStorage API for Array Integration (VAAI).You may use storage arrays that are certified for VAAI only with VMware vSphere Enterprise, VMware vSphere Enterprise Plus, VMware vSphere with Operations Management Enterprise, and VMware vSphere with Operations Management Enterprise Plus editions.

In the VMware EULA you can find the following:

3.1 License Restrictions. Without VMware’s prior written consent, You must not, and must not allow any third party to: (a) use Software in an application services provider, service bureau, or similar capacity for third parties, except that You may use the Software to deliver hosted services to Your Affiliates; (b) disclose to any third party the results of any benchmarking testing or comparative or competitive analyses of VMware’s Software done by or on behalf of You, except as specified in Section 2.4 (Benchmarking); (c) make available Software in any form to anyone other than Your employees or contractors reasonably acceptable to VMware and require access to use Software on behalf of You in a matter permitted by this EULA, except as specified in Section 2.2 (Third Party Agents); (d) transfer or sublicense Software or Documentation to an Affiliate or any third party, except as expressly permitted in Section 12.1 (Transfers; Assignment); (e) use Software in conflict with the terms and restrictions of the Software’s licensing model and other requirements specified in Product Guide and/or VMware quote; (f) except to the extent permitted by applicable mandatory law, modify, translate, enhance, or create derivative works from the Software, or reverse engineer, decompile, or otherwise attempt to derive source code from the Software, except as specified in Section 3.2 (Decompilation); (g) remove any copyright or other proprietary notices on or in any copies of Software; or (h) violate or circumvent any technological restrictions within the Software or specified in this EULA, such as via software or services.

My tests and both of this excerpts brings me to the conclusion that if a customer doesn’t have an Enterprise or Enterprise Plus license he is not allowed to use VAAI and therefor should disable VAAI to not violate the EULA.

You can find the procedure to disable VAAI in this KB article.

97 total views, 3 views today

In Part 1 and Part 2 I covered low power designs and medium designs of my homelab design process. In this part I cover one high end design which was inspired by the builds from Erik Bussink and Frank Denneman. My build is mainly (80%) the same as the two other builds but with some modifications I will explain at the end of the post.

High End Design

Advantage

• Xeon E5-2600 Processor
• 6 Core CPU
• Hyper-Threading (12 threads)
• 64GB memory
• two X540-AT2 10Gbit NICs
• IPMI (incl. KVM)
• Hardware RAID Controller LSI2308
• expandable (HDD)
• VT-x/VT-d
• complete HCL supported
• Power consumption compared to Performance

Disadvantage

• Price

So what is different to the other builds and why have I chosen it?

CPU

Erik and Frank have chosen the E5-1650 v2 processor which gives them much higher clock speed compared to the E5-2620 v2 (3.5GHz/3.9GHz TurboMode vs 2.1GHz/2.6GHz). But there is also a drawback with higher clock speed which is TDP (130Watts vs 80Watts). This can lead to a higher power consumption which should be low for my requirements. The most significant drawback from my point was the price. The E5-1650 v2 is approx. 45% more expensive than the E5-2620 v2 and the only benefit is clock speed (performance). One of the benefits of the E5-2620 v2 is, that you can use it in a dual socket mainboard. One problem encountered by Reinhard Partmann was when he tried to use 32GB LRDIMMs with the E5-1650 v2, it didn’t work. His reseller told him that this DIMMs were not supported with the E5-1650 v2. I tried to find some documentations from Intel for this problem but until now I didn’t find anything. When I have new information about that I will update this post.

Memory

Samsung RDIMMs were the only available DIMMs at the time of buying at my reseller, so I bought this instead of the Kingston RDIMMs.

SSD

Because the whole price of the server was one of my requirements, I have chosen Samsung 840 EVO SSD instead of Enterprise SSDs like the S3700.

In Part 4 of this series I will do a quick summary across all designs in terms of price and present my final build.

Update:

I sent some mails to SuperMicro and Intel regarding the problem with the E5-1600 processor and LR-DIMMs and got the following back from Intel customer support:

While the usage of LRDIMM modules are limited for Intel® E5-1600 processors v1, LRDIMM modules will no longer supported since Intel® E5-1600 processor v2. And they will not be referred anymore at Intel® E5-1600 processors v3.

You can find the required information in the datasheet of the E5-1600 v1 (page 15, section 1.2.1 System Memory Support) and of the E5-1600 v2 (page 17, section 1.2.1 System Memory Support). In the datasheet from the E5-1600 v3, section System Memory Support was removed.

LRDIMM x4, x8 (2-Gb and 4-Gb only) LR-DIMMs are supported only on server specific processors (Intel® Xeon® processor E5-2600 v2 and E5-4600 v2 product families). LR-DIMMs are not supported in workstation specific SKUs such as the Intel® Xeon® processor E5-1600 v2 product family.

296 total views, 1 views today