I am running several standalone versions of esxi with 4-5 vms on a 1TB local datastore. I also have an identical esxi server acting a a standby. I want to be able to copy or rsync the vmdk files after a snapshot to the standby esxi server. But I think this will take a very long time and not sure about how I should schedule the backup schedule.
Is it possibly to sync only the difference in vmdk files. I know veeam backup and replication can do that but I am looking for a free low cost solution for this.
what do you think?
The ability to capture changes requires access to the storage APIs which are not available in the free versions of ESXi. None of the commercial backup software that use change block tracking will work with the free versions of ESXi. You can have a look at ghettoVCB which can clone your VMs to another location. http://communities.vmware.com/docs/DOC-8760 You would need to provide a destination storage device such as an NFS device which could be shared by both ESXi hosts.
i have a 2008 NFS server running anf it taking a long time for the compressing part. Do you think its better to have NFS running on a linux vm or windows?
thanks
IMHO I prefer using Linux for NFS server.
Be sure to enable jumbo frames to have better performance.
Andre
have a look at this: http://www.openfiler.com/
i have a 2008 NFS server running anf it taking a long time for the compressing part. Do you think its better to have NFS running on a linux vm or windows?
What is taking a long time to compress? Where is the compression taking place? ESXi or NFS server?
i got another issue. I can only create a 256GB NFS Disk max. I guess this has to do with the 1MB block size. Since this is on localdisk, how do I change it from 1MB to 4MB?
You would need to change the block size on the datastore. That means migrating all the Virtual machines and recreating the datastore and then moving the VMs back to the datastore..
how do I create a RDM on a local datastore? Is that possible?
You can use a process like this - http://www.vm-help.com/esx40i/SATA_RDMs.php
I am getting this error
i backed up a vm to nfs and took 1380minutes to backup 280gb vm
anything i can do to speed this up
Make sure your network is GB from end to end. Check network duplex settings. Caching RAID controllers and multiple high speed disks in your RAID array on the ESXi host and NFS server.
its 100mb only and the nfs server is a vm not rdm. but when i backed up a 80gb vm it only took 40mins
We can't see how you have you infrastructure set up so it is difficult to give you specific help. You will need to go through things and elminate problems. 100MB network is a problem. The time taken to back up 280 GB looks to be excessive but not necessarily if it is going across a 100MB connection. Is the 80GB backup a thin disk to a thin disk where the actual used space was actually much less than 80GB?
for the 280GB vm, they are all thin disks. for the 80GB vm, it has both thin and thick disks.
Like I said there isn't much anyone can say to help. You have a unique setup and you will need to discover the issues. Testing and viewing logs etc. I would at the very least get everything on GB connections.