Question

Space really available in a container with thick provisioned VM

  • 22 January 2019
  • 2 replies
  • 1550 views
S
Hi all,

we are performing a POC and we have a cluster that originallly had about 30 Tib of available space made of SSD+ HDD. Then, to show the customer the performance improvement moving to All Flash we made an inline conversion addind SSD and removing all the HDD and now we have a total available space is about 11 TiB.

存储容器具有压缩启用，实际使用的容量为4,5 TIB。

The problem is that most of the VM are thick provisioned and the actual use space is reported at 12 Tib hence above the available capacity. This triggers RED alarms on the Prism console saying that he Data Resiliency is not available.

So I have 2 questions:

1）由于实际使用的容量为4,5 TIB，如果磁盘故障发生了什么？
2) how can we manage to eliminate these alarms?


Question number 2) is particularly relevant since if we succeed in the POC and the customer could purchase Nutanix all flash systems , they will like to use them without changing their policies and exploitining the data reduction functionalities of Nutanix.
.
Today they have many VMware cluster with Pure Storage systems that use LUNs compressed and deduped in the background without "complaining" about thick provisioning...

Thanks in advance for any suggestion.

Stefano


1个附件
    \r\n
    \r\nwe are performing a POC and we have a cluster that originallly had about 30 Tib of available space made of SSD+ HDD. Then, to show the customer the performance improvement moving to All Flash we made an inline conversion addind SSD and removing all the HDD and now we have a total available space is about 11 TiB.
    \r\n
    \r\nThe storage container has compression enable and the real used capacity is 4,5 Tib.
    \r\n
    \r\nThe problem is that most of the VM are thick provisioned and the actual use space is reported at 12 Tib hence above the available capacity. This triggers RED alarms on the Prism console saying that he Data Resiliency is not available.
    \r\n
    \r\nSo I have 2 questions:
    \r\n
    \r\n1) since the actual used capacity is 4,5 Tib what happens in case of a disk failure?
    \r\n2) how can we manage to eliminate these alarms?
    \r\n
    \r\n
    \r\nQuestion number 2) is particularly relevant since if we succeed in the POC and the customer could purchase Nutanix all flash systems , they will like to use them without changing their policies and exploitining the data reduction functionalities of Nutanix.
    \r\n.
    \r\nToday they have many VMware cluster with Pure Storage systems that use LUNs compressed and deduped in the background without \"complaining\" about thick provisioning...
    \r\n
    \r\nThanks in advance for any suggestion.
    \r\n
    \r\nStefano
    \r\n
    \r\n
    \r\n

    <\/a><\/p>","quoteUsername":"Stefano P.","translations":{"Common":{"like":"Like","unlike":"Unlike"},"Forum":{"Quote":"Quote","Share":"Share"}}}">

    This topic has been closed for comments

    2 replies

    manfred
    Userlevel 2
    Badge +7
    as nutanix uses NFS as protocol and supports the VAAI functions. therefore you can deploy thick provisioned VMs to a NFS datastore (which is not possible without VAAI).
    Nutanix ADSF reserves the blocks for thick provisioned VMs; (as you requested by your VM deployment)
    but for efficiency and storage capacity purposes it is better to go for thin; not only for Nutanix but also for Pure: both platforms support thick VMs but both recommend thin provisioning:

    http://www.joshodgers.com/2017/07/27/its-2017-lets-review-thick-vs-thin-provisioning/

    https://support.purestorage.com/Solutions/VMware_Platform_Guide/001VMwareBestPractices/PDF_Guide%3A_VMware_vSphere_Best_Practices_for_the_Pure_Storage%C2%AE_FlashArray
    \r\nNutanix ADSF reserves the blocks for thick provisioned VMs; (as you requested by your VM deployment)
    \r\nbut for efficiency and storage capacity purposes it is better to go for thin; not only for Nutanix but also for Pure: both platforms support thick VMs but both recommend thin provisioning:
    \r\n
    \r\nhttp:\/\/www.joshodgers.com\/2017\/07\/27\/its-2017-lets-review-thick-vs-thin-provisioning\/
    \r\n
    \r\nhttps:\/\/support.purestorage.com\/Solutions\/VMware_Platform_Guide\/001VMwareBestPractices\/PDF_Guide%3A_VMware_vSphere_Best_Practices_for_the_Pure_Storage%C2%AE_FlashArray","quoteUsername":"manfred","translations":{"Common":{"like":"Like","unlike":"Unlike"},"Forum":{"Quote":"Quote","Share":"Share"}}}">
    manfred
    Userlevel 2
    Badge +7
    in regard to data resiliency;
    reserved blocks are treated as if there are in use: so they cannot be used for example as failover capacity. if you take a look on the data resiliency in detail (click on data resiliency in prism home screen) you see if you can loose a disk or not.(FAULT DOMAIN TYPE: DISK)
    \r\nreserved blocks are treated as if there are in use: so they cannot be used for example as failover capacity. if you take a look on the data resiliency in detail (click on data resiliency in prism home screen) you see if you can loose a disk or not.(FAULT DOMAIN TYPE: DISK)","quoteUsername":"manfred","translations":{"Common":{"like":"Like","unlike":"Unlike"},"Forum":{"Quote":"Quote","Share":"Share"}}}">
    Terms & Conditions
    Learn more about our cookies.<\/a>","cookiepolicy.button":"Accept cookies","cookiepolicy.button.deny":"Deny all","cookiepolicy.link":"Cookie settings","cookiepolicy.modal.title":"Cookie settings","cookiepolicy.modal.content":"We use 3 different kinds of cookies. You can choose which cookies you want to accept. We need basic cookies to make this site work, therefore these are the minimum you can select. Learn more about our cookies.<\/a>","cookiepolicy.modal.level1":"Basic
    Functional","cookiepolicy.modal.level2":"Normal
    Functional + analytics","cookiepolicy.modal.level3":"Complete
    Functional + analytics + social media + embedded videos"}}}">
    Baidu