Around the Storage Block Blog
Find out about all things data storage from Around the Storage Block at HP Communities.

The new vSphere 5.0 VAAI primitives

CartoonCalvin100X100.JPG By Calvin Zito, @HPStorageGuy vmw-vexpert.jpg

 

I have one more blog post for you today from the VMware launch.  This last one is on a subject near and dear to me, VAAI.  Just yesterday (Monday July 11) I had a conversation with Kiran Sreenivasamurthy, one of our HP 3PAR product managers, about four of the new VAAI primitives and I'll share that podcast with you shortly.

 

Before I mention what the new primitives are, I want to tell you that I'm seeing multiple names for some of these.   In the podcast with Kiran, we were using the names in the engineering docs from VMware.  I'm now reading a VMware technical marketing paper called "What's New in vSphere 5.0 - Storage" and the names they're using there are a bit different than the namesKiran and I used in the podcast.  Here's briefly what those primitives are - first with what Kiran and I used in the podcast followed by the name I see in the white paper:

 

  • Unmap (called Dead Space Reclamation in white paper): Used for space reclamation rather than WRITE_SAME. Reclaim space after a VMDK is deleted within the VMFS environment.  And as you'll hear Kiran and I discuss, we think we have some advantages here with the HP 3PAR architecture.
  • TP LUN Reporting (general discussed in the white paper as "out of space condition"): notifies whether or not a LUN is a Thin Provisioned LUN without going back to a vendor's management tool to check.
  • Out of Space Condition (general discussed in the white paper as "out of space condition"): Uses CHECK CONDITION status with either NOT READY or DATA PROTECT sense condition - in other words, instead of having a VM blue screen when it runs out of space, VMware pauses to provide enough time so that the associated volumes can be expanded to proceed.
  • Quota Exceeded Behavior (general discussed in the white paper as "out of space condition"): VMware provides alerts when certain thresholds are exceeded.  These thresholds are set with the storage vendor's tools. 

 

Kiran and I didn't talk about the new primitives for NAS but you can read about them in the VMware paper. I wanted to attach it to my post but the file is too big.  Here's a link to the VMware paper .  

 

Here's the podcast with Kiran - while I've briefly written about some of the new primitives, I'd encourage you to listen to the podcast as Kiran and I covered not only the new primitives but talked about how HP 3PAR takes advantage of these and some of the vSphere 4.1 primitives. 

 

 

Here's a link where you can download the podcast for your future listening pleasure on your MP3 player of choice.

 

If you haven't seen the previous vlogs I did on VAAI with the P4000, those are recommended viewing to help you better understand VAAI.  I did a two part series, first giving a white board overview of VAAI:

 

 

... and part two was an actual demo of me showing the XCOPY primitive in action - doing a clone and vMotion with a P4800.  Here's that video:

 

 

Come on back tomorrow as I still have three more podcasts to share with you - SRM 5.0, VSA, and 2TB memory support in a server. Now hopefully I can have them all completed by tomorrow! 

Labels: podcast| storage| VAAI| VMware
Comments
chriswong13(anon) | ‎07-13-2011 07:28 AM

Don't need to see the 2 TB memory demo.  Who's gonna pay over $ $150,000 in Enterprise Plus licenses on top of server hardware costs to run 2 TB of RAM in a server...?  Ridiculous...

| ‎07-13-2011 07:48 AM

Hey Chris - the outcry on this has been loud.  While it certainly isn't my job to defend VMware's pricing, I'm talking with my server team to understand this better.  I'll hopefully have more tomorrow.  But, I hear you!

| ‎07-13-2011 05:31 PM

After writing this post, I found a VMware blog that gives a very good overview of the new VAAI primitives.  Here's a link to that article: http://blogs.vmware.com/vsphere/2011/07/new-enhanced-vsphere-50-storage-features-part-3-vaai.html

Derek(anon) | ‎07-14-2011 01:49 AM

So when will 3PAR and the P4000 support the new VAAI commands?

Derek(anon) | ‎07-14-2011 03:25 AM

 

I performed several tests of the VAAI 1.0 extensions with 3PAR, and got really good results. I look forward to the VAAI 2.0 support and performing more tests.

 

http://derek858.blogspot.com/2010/12/3par-vaai-write-same-test-results-upto.html

http://derek858.blogspot.com/2010/12/3par-vaai-xcopy-vmotion-test-results.html

 

| ‎07-14-2011 07:49 AM

Hi Derek - not sure what I can say publicly about when HP 3PAR and P4000 will support the latest VAAI primitives in vSphere 5.0.  HP doesn't usually talk futures in public.  You should be able to ask your HP rep and they can tell you under NDA.  I'll see what I can say and let you know.  Also, thanks for the links.  I'll definitely check them out.  I haven't seen VAAI on HP 3PAR myself yet - only on the P4000. 

PEDRO J(anon) | ‎10-01-2012 06:37 PM

Hello!

 

I have a 3par F200 whit  Inform OS 3.1.1 and ESX 5.

 

The manual says, no need any pluig in for this versions, but when delete any vmware machine on 3par THIN (zero detect option on) datastore,  the size of the volume is the same.

 

When copy any vmachine from datastore full to thin 3PAR it´s ok all, the size are small than the original on full volume.

 

When delete the zeros (space reclamation) after delete de virtual machine (vmdk) on datastore wmfs 5?

 

Best reguards.

| ‎10-02-2012 12:02 AM

Pedro,

 

I think you're saying that when you reclaim space after deleting a VMDK you don't see a space decrease - is that right? If yes then here are a few things:

  1. What exact version of ESXi5 is in use.  ESXi5 patch2, ESXi5U1 and ESXi5.1 will not reclaim on the fly after a vmdk is deleted.  You must use the vmkfstools –y X command after deleting the VMDK.  Where X is a value between 1-100%.
  2. You can run esxtop, add the VAAI counters and ensure that block delete is incrementing when you run the vmkfstools command.
  3. The space reclaimed is not exact when you use vmkfstools.  That is because some of the space to be reclaimed may have never been allocated hence a user after running the commands will not notice any space     reclaimed on the array side.
  4. Also if you have zero detect on, you should create a eagerzeroedthick VMDK as a way to allocate space for the VMDK because no physical capacity will be allocated.  Deleting this VMDK will not amount to any space being reclaimed.

Hope that helps!

Leave a Comment

We encourage you to share your comments on this post. Comments are moderated and will be reviewed
and posted as promptly as possible during regular business hours

To ensure your comment is published, be sure to follow the community guidelines.

Be sure to enter a unique name. You can't reuse a name that's already in use.
Be sure to enter a unique email address. You can't reuse an email address that's already in use.
Type the characters you see in the picture above.Type the words you hear.
Search
Showing results for 
Search instead for 
Do you mean 
About the Author
25+ years experience around HP Storage. The go-to guy for news and views on all things storage..


Follow Us
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation