I don't think VCB is a joke at all. Especially when integrated with a product that fully integrates with VCB, like Arcserve 12. No longer do you have to script command line commands, nor do you have to deal with scripts and config files. A GUI is used to discover your VM's, and write them to the database. Then, they are listed in the column at the left just like any other network discovered machine. No pre and post scripts, no command line interfaces, all menu driven. I am sure all of the other major software backup brands will follow suit, if they haven't already. For us, this plus agent-driven backups for Exchange/Oracle/AD/etc. gives us a two pronged approach to backup, both fullvm bare-metal restore capability, and database backups that are not just crash consistent. And, as mentioned, it is fast.
If you found this or any other post helpful please consider the use of the Helpfull/Correct buttons to award points
This looks like a fantastic product. I am a big fan of Netapp and one of their other snaphot products SME. Thanks for sharing this link.
Yes it sounds like an interesting product, I should have a better view by the end of the day as I am off to NetApps to see it in action. the project I am currently working on uses NetApp technology and are lookig at DR/Business continuity at the moment.
Tom Howarth
VMware Communities User Moderator
Netapp did the announcement a couple of weeks ago (while we were in the process of buying our new storage).
Unfortunately our system integrator (netapp partner ofcourse) told us that Netapp hadnt published the prices yet so we couldnt add it in our project.
Btw they told us that probably the price will be aligned to other SM products. I'm waiting for news from them...we'll see...
Price is one of the questions I will be asking of them today
Tom Howarth
VMware Communities User Moderator
I will keep an eye on this and update my results, this is what I'm implementing within the next few weeks using NFS storage on the filers.
On H Class Blades
We use vRanger. The latest version (I think it came out a few weeks ago) is light-years ahead of the previous version. They fixed a lot of the scheduling issues.
I hear esXpress is good..... and FREE. you can't beat that.
VCB is a joke. It's a pain to setup. It's confusing to use. it doesn't always seem to work. When it does work, it's really fast. I'm talking a reduction in backup times by 70 to 80%. But, our backups across the network really don't take that long with Ranger. I suppose we'll hit the wal at some point where there aren't enough hours in the day to do backups and we'll HAVE to use VCB but until that happens, I'll just stick with network backups.
The NetApp stuff sounds very good. The only major snag I see is that the snapshot remain on the same disk spindles as the production VM which means you are not protected from a SAN/disk failure. You could of course replicate these off with Snapmirror but that gets pricey compared to the software based options.
of course this is not the solution of all the problems, but it allows you to recover from a single vmware / datastore crash within minutes knowing that the data will be vmware-consistent[*]
If you have other requirements, you can think about additional solution like backup via NDMP to a tape library (for offsite storage) or replication via SnapMirror (for DR purposes)
if you think that crash-consistent data are sufficient for you, you can schedule netapp's snapshot of vmware volumes without buying SM for VI (i used in the previous shop, and when we had to mount a snapshot and take out and old snap of a VM, it worked).
We had a netapp rep come and in explain the technology. Basically it's the same as their snap technology now, where you can take a snapshot of an entire LUN (provided you have sufficient reserve space on the volume) in 3 seconds. yes 3 seconds.
They took this technology, applied to VM Ware, and they claim to have some sort of driver that is ESX aware and take take a snapshot of individual VM's in a LUN in the same time frame.
I dont' care how fast your backups are now, taking a snapshot of a LUN or VMFS volume can be done in less than minute, at once, NO backup agents, no VCB proxy, no additional software needed on the ESX server side needed.
So this is instantaneous backups, driven by the SAN device, and restores are individual, not LUN level.. according to Netapp. We want to test this, but if this works, this will be the absolute best way to make backups bar none.
You can't compete with having backups on the SAN done by the SAN, and in under a minute. Restores same way.
Just realize that Netapp's SnapShot technology is not a true "snapshot" in the way we were familiar with. Netapp uses "pointers to deltas" so it does not take up huge disk space as compared to other snapshots. Very cool performance comparison in their snapshots when another systems uses snapshots compared to Netapp's. Their SnapManager products (Exchange and SQL) are great products which can work directly or work inconjuction with an enterprise backup solution like Symantec and CommVault.
I don't think VCB is a joke at all. Especially when integrated with a product that fully integrates with VCB, like Arcserve 12. No longer do you have to script command line commands, nor do you have to deal with scripts and config files. A GUI is used to discover your VM's, and write them to the database. Then, they are listed in the column at the left just like any other network discovered machine. No pre and post scripts, no command line interfaces, all menu driven. I am sure all of the other major software backup brands will follow suit, if they haven't already. For us, this plus agent-driven backups for Exchange/Oracle/AD/etc. gives us a two pronged approach to backup, both fullvm bare-metal restore capability, and database backups that are not just crash consistent. And, as mentioned, it is fast.
If you found this or any other post helpful please consider the use of the Helpfull/Correct buttons to award points
@rparker: i have some doubt about how they can assure you they can recover granularly at VM-level when you're snapshotting a FC/iSCSI LUN (that is a file in their WAFL filesystem) - maybe they do some kind of "transparent LUN mounting" and then extract the single VM files on the fly without user interaction ? mmm
On the contrary I can easily understand that it can works on NFS implementation (where machines resides directly on the volume without any additional layer, so the "only" addition of SM is communicating with VC/ESX in order to take an consistent snap)
@dmorgan: VCB is not a joke but it's a different thing: it's a storage-agnostic approach that allows you to "extract" the virtual machines files without "disturbing" ESX servers (before VCB you had to rely on ESX access to the LUNs in order to copy out the files). It's "agnosticism" is a pro but also a con: a pro because it's a generic approach usable in every FC/iSCSI situation for backup but it's a con because you are not using any advanced feature your storage vendor could offer you (in this case, netapp instantaneous snapshot capabilities for local or remote (DR) recovery).
Prolly the best approach is using snapshots (and Snap Manager for Virtual Infrastructure) for local immediate recovery (or for DR replication), and use VCB for tape backup and subsequent offsite vaulting if you cant afford a SAN tape directly connected to the Netapp storage.
My 2 cent (eurocent, because dollar is so weak )
I agree. We too have a Netapp SAN. Well, IBM re-branded, but Netapp nonetheless :). However, we use this SAN for more than just VMWare, and unfortunately we don't typically have the space to reserve enough for snapshots to be housed for DR. VCB, plus Arcserve agents gives us a good solution, if not the best solution. We can still do VM snaps for imediate needs, and VCB can do bare metal backups to be offloaded to the proxy fairly quickly. We may not be taking full advantage of the netapps abilities, but I can live with that, at least until I can convince the powers that be to add a couple more trays of disks!
-D-
If you found this or any other post helpful please consider the use of the Helpfull/Correct buttons to award points
Sound like a good solution for those using Netapps. I'll be interested to see how it really works.