2011-02-25 05:40 PM
In searching knowledge bases and manuals I have been unable to find any reference to why or how the copy offload option of mbralign would be used. Anyone know. The name and having to provide the login and nfs path to implement it gives some indication, but I'd rather not go on simple speculation based on that.
I've tried running some tests with a few syntaxes variations like the following:
/opt/netapp/santools/mbralign/mbralign --access 172.31.2.151:root:<password> --path /vol/cgy3a_t10_os01/cgyspsp20-testalign --sparse cgyspsp20-testalign_1-flat.vmdk
Based on the help info shown below I'm assuming for controller the controller IP would work
I'm more confused about the --path parameter. Just guessing that it may want the NFS path to the datastore (or VM directory, or specific vmdk file, or vmx file, or ...)?
MBRALIGN is happy enough with the above options to go as far as begining the backup:
Creating a backup of cgyspsp20-testalign_1.vmdk
Creating a backup of ./cgyspsp20-testalign_1-flat.vmdk
Creating a copy of the Master Boot Record
Working on partition P1 (3): Starting to migrate blocks from 32256 to 32768.
Error: <results status="failed" reason="Extra input: disable-hole-punch" errno="13115"></results>
All the info I can find on it is from 'mbralign --help':
Usage: mbralign [OPTION]... SOURCE
or: mbralign [OPTION]... SOURCE DEST
or: mbralign [OPTION]... SOURCE DIRECTORY
-a|--access controller:login:passwd Credentials for the controller, login, and password. Used for NFS Copyoffloading only.
--path The NFS path. Used for NFS Copyoffloading only.
--sparse The resulting file will be sparse. This option is
designed for NetApp NFS based datastores.
--force The default behavior is to skip properly aligned disks.
The --force option can be used to
override this. (Useful for migrating and making sparse).
--bs=n Use n as the block size (specified in kB). n must be one
of (8, 16, 32, 64, 128, 1024). The default is 8.
-s|--secure Use SSL for secure mode
--preview Show what would be done, then exit. No changes are made
to the file.
--scan all Show current alignment of all flat-vmdk files, then exit. No changes are made
filename Show alignment of one flat-vmdk file, then exit. No changes are made
--quiet Only print warnings and errors (no other status is
printed to screen).
--debug Print debug comments.
--help Print usage.
--thereAreNoVmSnapshots Suppress the warning message about snapshots and
answer YES to the question of whether to continue.
2011-02-28 01:38 PM
Support and documentation for the mbralign copy offload feature will be in the next release of mbralign. The new mbralign is included in the next version of ESX Host Utilities, which should be available from the NetApp Support site soon.
2011-03-03 12:57 PM
The ESX Host Utilities 5.2.1 software package is now available from the NetApp Support (NOW) site. It includes a new version of the mbralign program that supports copy offload.
For more information, see the topic "Offloading VMDK data copying improves performance for NFS" in the ESX Host Utilities 5.2.1 Installation and Setup Guide.
Note that the copy offload feature is available with storage controllers running Data ONTAP 7.3.5 and later or Data ONTAP 8.0.1 and later software.
2011-03-04 04:55 AM
Thanks for the info.
FYI: I'm receiving a 'the requested item could not be found' error when trying to download the file. The download link on the download page appears to have a normal download path to the file, https://now.netapp.com/NOW/download/software/sanhost_esx/5.2.1/netapp_esx_host_utiliti
I'm wondering if there is an error with the posting of the file, or if there is just a delay in how the download file is made available after the downlaod page is viewable. I'll try again later.
2011-03-17 09:19 PM
FYI for anyone looking at this new version copy offload performance. In initial tests of it I have been seeing alignments complete at a speed of 10 GB per minute. Much better than the performance without copy offload.
2012-08-02 12:39 PM
can you post an example of a working command line? I am having issues trying to get the offload option to work. In my case the datastore is hosted on a vFiler. The command fails with
NFS Copy offloading could not complete successfully.
2013-01-09 06:46 AM
I have the exact same problem.
I reverted to using HTTP and run a tcpdump while executing.
This lead me to the exact same issue I have when using the copyoffload command directly through the ZAPI:
<?xml version='1.0' encoding='UTF-8' ?>
<!DOCTYPE netapp SYSTEM '/na_admin/netapp_filer.dtd'>
<netapp version='1.1' xmlns='http://www.netapp.com/filer/admin'>
So, the NetApp box has a problem with that. Up to now I was unable to figure out what the problem might be.