UNSOLVED

nicolasm1

updated

14 years ago

N

nicolasm1

5 Posts

0

1903

July 5th, 2012 01:00

VNXe 3100 - incoherence datastore capacity on UNISPHERE vs on Vcenter

Hello everyone,

I have a VNXe3100

I don't understand why my capacity datasore is different between UNISPHERE and VCENTER

(see attachement)

The total size is correct 1.99 To but I have a problem with remaining space and so the current allocation

(1.05 To on Vcenter <> 0.03 To on Unisphere)

Is there a tool to defrag the datastore on vCenter or Unisphere (not in the guest OS) or a tool to refresh, purge the datastore ?

Thank you for your help

1 Attachment

  • Jim_Hegner

    2 Intern

    •

    212 Posts

    600

    1

    Posted July 5th, 2012 01:00

    Hi

    We have seen something Similar before...

    From your screen shot is seems that you are running Vsphere Client on a Host with French language, and I will assume that your regional settings are also French...

    According to EMC support the VSI plugin does not support any other Regional settings than US ENGLISH.

    The problem is when the VSI plugin sends off NAVISECCLI commands to the Storage array something goes wrong with the way the Decimal symbols are interpreted....

    I'm working in Denmark and we cannot use the Vsphere client or VSI on a system with Danish regional settings.

    This is what we got from EMC engineering December last year

    The systems that cannot do the extend have a system locale of da:Danish as seen from the log. This changed the way we interprete the output from the NaviSecCLI, where the storage system locale is en_US. Specifically a decimal symbol in en_US:English is interpreted as a digit grouping symbol in da:Danish. This resulted in a wrong lun size and later wrong available space on the extent. You may also noticed a relative bigger number shown on the extend storage dialog.

    Since VSI doesn't officially support any local other than US English, and the workaround has satisfied the customer's need we need to close the eng ticket.

    Hope this Helps...

  • Sriky

    1 Rookie

    •

    45 Posts

    600

    1

    Posted July 5th, 2012 13:00

    Hi,

    This datastore is thin provisioned.

    VNXe will only see the raw space and if the ESX server had expanded to the maximum size of the datastore and if you delete some VM's or files on the VMFS datastores VNXe will not show the freed up space on the VMFS.

    Example: Initially a thin LUN with max size of 100gb  is created and given to the ESX server and by default its 10Gb which is allocated to the ESX and when you create 50Gb VMDK on the ESX now the VNXe will show 50gb usage and when you delete this 50Gb VMDK on the ESX, VNXe will not show the entire free space and it still continues to show 50gb usage.

    Reclaimation is the feature which comes with VAAI which reclaims the space and provision it back to the storage as of now VAAI for iSCSI is not supported.

    Regards,


    Sri

  • nicolasm1

    5 Posts

    600

    0

    Posted July 9th, 2012 05:00

    Thanks for your answers.

    But How can i "force" refresh to have the same capacity on the 2 sides ??

    THX

  • Sriky

    1 Rookie

    •

    45 Posts

    600

    0

    Posted July 9th, 2012 16:00

    There is no command or option to force refresh, you will have to keep track of the usage on the VMFS or use thick datastores.

    Regards,

    Sri